You are on page 1of 102

Centre de recherche sur l’environnement alpin (CREALP

)
HydroCosmos SA

RS MINERVE

RS MINERVE - Technical Manual v1.14
April 2015
For Software version 1.3.1.0

Approved for Public Release. Distribution Unlimited

Centre de recherche sur l’environnement alpin (CREALP)

HydroCosmos SA

Authors:

Javier GARCIA HERNANDEZ

CREALP

Javier PAREDES ARQUIOLA

UPV

Alain FOEHN
Bastien ROQUIER

CREALP
HydroCosmos S.A.

Please refer this technical manual as :
García Hernández, J., Paredes Arquiola, J., Foehn, A. and Roquier, B. (2015). RS MINERVE – Technical manual v1.14. RS
MINERVE Group, Switzerland.

Index

Index
Index ........................................................................................................................................... 4
Foreword .................................................................................................................................... 6
Chapter 1: Introduction.......................................................................................................... 7
1.1 Manual’s structure ...................................................................................................... 7
Chapter 2: Hydrological models description .......................................................................... 8
2.1 List of objects ............................................................................................................... 8
2.2 Virtual Weather Station ............................................................................................... 9
Thiessen interpolation ........................................................................................................ 9
Shepard interpolation ....................................................................................................... 10
Complementary calculations for the Potential Evapotranspiration (ETP) ....................... 11
2.3 Snow-GSM model description ................................................................................... 14
2.4 Glacier-GSM model description................................................................................. 16
2.5 Infiltration (GR3) model description.......................................................................... 18
2.6 Runoff (SWMM) model description .......................................................................... 20
2.7 GSM model description ............................................................................................. 21
2.8 SOCONT model description ....................................................................................... 24
2.9 HBV model description .............................................................................................. 28
2.10 GR4J model description ............................................................................................. 32
2.11 SAC-SMA model description ...................................................................................... 36
2.12 Channel routing description ...................................................................................... 45
2.13 Reservoir .................................................................................................................... 50
2.14 Level-Discharge Relation (HQ) ................................................................................... 51
2.15 Hydropower ............................................................................................................... 52
2.16 Diversion .................................................................................................................... 54
2.17 Consumer................................................................................................................... 55
2.18 Structure efficiency.................................................................................................... 56
2.20 Time series ................................................................................................................. 57
2.21 Sensor ........................................................................................................................ 58
2.22 Regulation and State ................................................................................................. 59
Chapter 3: Performance indicators ...................................................................................... 60
3.1 Nash coefficient ......................................................................................................... 60
3.2 Nash coefficient for logarithm values ....................................................................... 60
3.3 Pearson Correlation Coefficient ................................................................................ 61
3.4 Bias Score ................................................................................................................... 61
3.5 Relative Root Mean Square Error .............................................................................. 61
3.6 Relative Volume Bias ................................................................................................. 62
3.7 Normalized Peak Error ............................................................................................... 62
RS MINERVE – Technical Manual

Page 4/94

Index

Chapter 4: Calibration of hydrological models .................................................................... 63
4.1 Introduction to the calibration module ..................................................................... 63
4.2 Objective function ..................................................................................................... 63
4.3 Shuffled Complex Evoluation – University of Arizona ............................................... 64
Model architecture ........................................................................................................... 64
Algorithm parameters ....................................................................................................... 69
4.4 Uniform Adaptive Monte Carlo ................................................................................. 69
Model architecture ........................................................................................................... 69
Algorithm parameters ....................................................................................................... 70
4.5 Coupled Latin Hypercube and Rosenbrock ............................................................... 71
Model architecture ........................................................................................................... 71
Algorithm parameters ....................................................................................................... 74
Chapter 5: Files formats ....................................................................................................... 76
5.1 Model files ................................................................................................................. 76
5.2 Parameters and initial conditions files ...................................................................... 76
5.3 Objects selection files ................................................................................................ 77
5.4 Exported results files ................................................................................................. 77
5.5 Calibration configuration files ................................................................................... 77
5.6 Database files............................................................................................................. 77
Chapter 6: Database formats ............................................................................................... 78
6.1 The RS Database Viewer ............................................................................................ 78
6.2 Database and dataset formats .................................................................................. 80
Database files .................................................................................................................... 80
Dataset files....................................................................................................................... 83
6.3 Automatic creation of databases or datasets ........................................................... 86
Chapter 7: GIS formats ......................................................................................................... 89
7.1 Model files ................................................................................................................. 89
Bibliography.............................................................................................................................. 90
Acknowledgments .................................................................................................................... 94
Appendix A:
List of notifications from pre-simulation and post-simulation reports......... A.1
A.1 Any isolated objects.................................................................................................. A.1
A.2 Base objects .............................................................................................................. A.1
A.3 Structures objects ..................................................................................................... A.4
A.4 Standard objects ....................................................................................................... A.7
A.5 Regulation objects .................................................................................................... A.8

RS MINERVE – Technical Manual

Page 5/94

Foreword

Foreword
RS MINERVE is a software for the simulation of free surface run-off flow formation and
propagation. It models complex hydrological and hydraulic networks according to a semidistributed conceptual scheme. In addition to particular hydrological processes such as
snowmelt, glacier melt, surface and underground flow, hydraulic control elements (e.g.
gates, spillways, diversions, junctions, turbines and pumps) are also included.
The global analysis of a hydrologic-hydraulic network is essential in numerous decisionmaking situations such as the management or planning of water resources, the optimization
of hydropower plant operations, the design and regulation of spillways or the development
of appropriate flood protection concepts. RS MINERVE makes such analyses accessible to a
broad public through its user-friendly interface and its valuable possibilities. In addition,
thanks to its modular framework, the software can be developed and adapted to specific
needs or issues.
RS MINERVE contains different hydrological models for rainfall-runoff, such as GSM,
SOCONT, SAC-SMA, GR4J and HBV. The combination of hydraulic structure models
(reservoirs, turbines, spillways,…) can also reproduce complex hydropower schemes. In
addition, a hydropower model computes the net height and the linear pressure losses,
providing energy production values and total income based on the turbine performance and
on the sale price of energy. A consumption model calculates water deficits for consumptive
uses of cities, industries and/or agriculture. A structure efficiency model computes
discharge losses in a structure such a canal or a pipe by considering a simple efficiency
coefficient.
The RS Expert module, specifically created for research or complex studies, enables in-depth
evaluation of hydrologic and hydraulic results. Time-slice simulation facilitates the analysis
of large data sets without overloading the computer memory. Scenario simulation
introduced the possibility of simulating multiple weather scenarios or several sets of
parameters and initial conditions to study the variability and sensitivity of the model results.
The automatic calibration with different algorithms, such as the SCE-UA, calculates the best
set of hydrological parameters depending on a user-defined objective function.
RS MINERVE program is freely distributed to interested users. Several projects and theses
have used and are using this program for study basins in Switzerland, Spain, Peru, Brazil
France and Nepal. In addition to the research center CREALP and the engineering office
HydroCosmos SA, which currently develop RS MINERVE, two universities (Ecole
Polytechnique Fédérale de Lausanne and Universitat Politècnica de València) collaborate to
improve RS MINERVE and use it to support postgraduate courses in Civil Engineering and
Environmental Sciences. Other collaborations, such as with the Hydro10 Association,
complement and enhance the development of RS MINERVE.

RS MINERVE – Technical Manual

Page 6/94

Chapter 1: Introduction

Chapter 1: Introduction
The Routing System II program was developed at the Laboratory of Hydraulic Constructions
(LCH) at the Ecole Polytechnique Fédérale de Lausanne (EPFL) (Dubois et al., 2000; García
Hernández et al., 2007).
The program presented hereafter, RS MINERVE, is based on the same concept than Routing
System II. RS MINERVE is developed by the CREALP and HydroCosmos SA with the
collaboration of the Laboratory of Hydraulic Constructions (LCH) at the Ecole Polytechnique
Fédérale de Lausanne (EPFL) and the Universitat Politècnica de València (UPV).

1.1 Manual’s structure
The presented Technical manual is organised in eight chapters:
1.
2.
3.
4.
5.
6.
7.

Introduction
Hydrological models description
Performance indicators
Calibration of hydrological models
Files formats
Database formats
GIS formats

For the RS MINERVE software utilisation, the reader can also use the RS MINERVE User’s
Manual (Foehn et al., 2015).

RS MINERVE – Technical Manual

Page 7/94

10)  SAC (SACRAMENTO-SOIL MOISTURE ACCOUNT) model description (see 2.5)  Runoff (SWMM) model description (see 2. but only a link to the database)  Comparator (object without any parameters or paired data)  Sub-model (object without any parameters or paired data)  Group Interface (object without any parameters or paired data)  Regulation objects:  Sensors (see 5. but only a link to another object for obtaining QWanted)  TurbineDB (object without parameters or paired data.17)  Standard objects:  Time Series (see 5.9)  GR4J model description (see 2.4)  Infiltration (GR3) model description (see 2.7)  SOCONT model description (see 2.16)  Consumer (see 2.  Base objects:  Virtual Weather Station (see 2.15)  Diversion (see 2. Not presented objects do not require data or require data from another object or from database. 2.19) RS MINERVE – Technical Manual Page 8/94 .2)  Snow-GSM model description (see 2.13)  Level-Discharge relation HQ (see 2.Chapter 2: Hydrological models description Chapter 2: Hydrological models description This chapter presents the hydrological objects existing in RS MINERVE software.6)  GSM model description (see 2.1 List of objects The parameters or required data (such as paired data) of all objects are presented hereafter.14)  Turbine (object without parameters or paired data.3)  Glacier-GSM model description (see 2.12)  Junction (object without parameters or paired data)  Structures objects:  Reservoir (see 2.8)  HBV model description (see 2.18)  Regulation and States (see 5.11)  Channel routing description (see 2. but only a link to the database)  Hydropower (see 2.18)  Source (object without parameters or paired data.

of Minimal number of stations used for stations interpolation (higher priority than “Search Radius”) Gradient P m/s/m Precipitation gradient Gradient T °C/m Temperature gradient Regular Range >0 ≥1 -a -0.3.2) Page 9/94 . min.5 to 2 Coeff T °C Adding correction coefficient -2 to 2 Coeff ETP Multiplying correction coefficient 0. temperature (Gradient T. This method has been extended to take into account the evolution of certain meteorological variables as a function of the altitude. Their regular ranges have to be estimated for each studied case. variables Pk. Z) allows the spatial distribution of the meteorological variables (precipitation. Station The method chosen for the spatial distribution of the precipitation. The second one. Table 1 List of parameters and initial conditions for the virtual meteorological station Object Name Units Description X.n is obtained by searching the nearest meteorological station k of the database to the virtual point s (normally referring to the gravity centre of a sub-catchment). dETP/dz).004 Gradient ETP m/s/m Evapotranspiration gradient -a Coeff P Multiplying correction coefficient 0. with spatial reference in metric coordinates. searches i stations being in a search radius and calculates the meteorological variable depending on inverse distance weighting. Additionally. PS  CoeffPs  ( Pk  dPs  ( z s  zi )) dz TS  CoeffT s  (Tk  dTs  ( z s  zi )) dz RS MINERVE – Technical Manual (A. the temperature and the ETP corresponds to the Thiessen and Shepard methods. potential evapotranspiration or temperature is also included in the method for obtaining the final value at virtual station s. temperature. Tk and ETPk. searches the nearest meteorological station for each meteorological variable.Chapter 2: Hydrological models description 2.007 to 0. Thiessen interpolation The evaluation of a variable in a virtual station s from n meteorological stations localized at i=1. Y. ETP) from available measures or estimations of a database.1) (A. This gradient dXs/dz corresponds to the gradients of precipitation (Gradient P. Shepard. dPs/dz). dT/dz) or potential evapotranspiration (Gradient ETP. Y. Z Coordinates of the virtual station Search Radius m Search radius of weather stations No. associated to the calculation point s. Thus. The first method.2 Virtual Weather Station The object « Virtual Station » (which is associated with the coordinates X. Thiessen. according to a linear relation with a constant elevation gradient dX s/dz.5 to 2 a The precipitation and evapotranspiration gradients are function of the local conditions. localized at the altitude zk are replaced by their equivalents related to the altitude of the virtual station s.2.1 to A. as presented in equations A.…. a coefficient for precipitation.

dT/dz) or potential evapotranspiration (Gradient ETP. In that case. temperature (Gradient T.2. localized at the altitude zi are replaced by their equivalents related to the altitude of the virtual station s.s. d i . yi: position of the meteorological station i of the database [m]. zs: altitude of the virtual station s [m a. s T  i 1 i dTs 1  ( z s  zi )  2 dz d i . ETPs: value of the potential evapotranspiration in the virtual station s [I.].4) with xi. Ts: value of the temperature in the virtual station s [I.7) Page 10/94 .Chapter 2: Hydrological models description ETPS  CoeffETPs  ( ETPk  dETPs  ( z s  zi )) dz (A. variables Pi.]. Ti and ETPi.n is obtained by weighting according to the inverse square distance d i. The n meteorological stations for the spatial interpolation in the virtual station s are determined automatically respecting equation A.s.].s between the meteorological station i of the database and the virtual station s. di.U. s RS MINERVE – Technical Manual (A. CoeffTs: temperature coefficient [°C].s n 1  2 i 1 d i . ys: position of the virtual station s [m]. since only the nearest meteorological station is used.6 to A. min.U.….U. dETP/dz). of stations (minimal number of stations) are not used. s  ( xi  x s ) 2  ( y i  y s ) 2 (A.s: distance between the meteorological station i and the virtual station s [m]. Shepard interpolation The evaluation of a variable in a virtual station s from n meteorological stations localized at i=1. Thus. Pk: value of the precipitation in the meteorological station i [I.U.l.3) with Ps: value of the precipitation in the virtual station s [I. Nevertheless. corresponding to the gradients of precipitation (Gradient P. as presented in equations A. ETPk: value of the potential evapotranspiration in the meteorological station i [I. rs). a minimal number of stations used for interpolation can be fixed by the user whit the corresponding parameter.l. Additionally. according to a linear relation with a constant elevation gradient dX s/dz.s n 1  2 i 1 d i . Hence. zk: position of the meteorological station i of the database [m a. Tk: value of the temperature in the meteorological station i [I.].8. a coefficient for precipitation.6) (A. potential evapotranspiration or temperature is also included in the method for obtaining the final value at virtual station s.U.5. CoeffPs: precipitation coefficient [-].5) with rs: search radius of meteorological stations [m]. d i .s  rs (A. the number n of meteorological stations is variable for every pair (s.]. n PS  CoeffPs  P  i 1 i n TS  CoeffT s  dPs 1  ( z s  zi )  2 dz d i . the parameters rs (search radius) and No.]. xs.U. The Shepard method (1968) has been also extended to take into account the evolution of the meteorological variables as a function of the altitude. dPs/dz).]. CoeffETPs: potential evapotranspiration coefficient [-].].

1961) is presented in equation A.4 K  0.9) with ETP : potential evapotranspiration [mm/month] . These methods can be selected in the RS MINERVE Settings.]. Complementary calculations for the Potential Evapotranspiration (ETP) If no ETP values are available in the database. zs: altitude of the virtual station s [m a. CoeffETPs: potential evapotranspiration coefficient [-].s. This data comes as a grid (latitude and longitude) and is composed of monthly averaged values.Chapter 2: Hydrological models description n ETPS  CoeffETPs   ETP  i 1 i dETPs 1  ( z s  zi )  2 dz d i .larc. ETPi: value of the potential evapotranspiration in the meteorological station i [I. The global radiation Rg is obtained in [kWh/m2/day] from the Global horizontal radiation dataset provided by the Surface meteorological and Solar Energy (SSE) web portal. a) Turc The potential evapotranspiration proposed by Turc (1955.]. Pi: value of the precipitation in the meteorological station i [I. K : constant [-].U.U.8) with Ps: value of the precipitation in the virtual station s [I.].37 if Month  February if Month February (A.s.]. T : air temperature [°C] . Figure 1 Selection of the ETP calculation The available methods are presented hereafter in detail.U.l.10) Rg value is a location dependent (latitude and longitude) monthly average of the global radiation. Ts: value of the temperature in the virtual station s [I.U.].l. on the Evapotranspiration frame (Figure 1).].U. CoeffPs: precipitation coefficient [-].nasa.U.].s n 1  2 i 1 d i . s (A. Rg : global radiation [cal/cm2/day] . The constant K value is: K  0. RS MINERVE offers also the possibility of calculating the ETP from different methods directly at virtual station. CoeffTs: temperature coefficient [°C]. sponsored by the NASA’s Applied Science Program (http://eosweb.]. RS MINERVE – Technical Manual Page 11/94 . ETPs: value of the potential evapotranspiration in the virtual station s [I.9: ETP  CoeffETP  K  ETP  0 T  ( Rg  50) if T  0 T  15 if T  0 (A. Ti: value of the temperature in the meteorological station i [I. zi: position of the meteorological station i of the database [m a.gov/sse).

on the basis of a study realized on many worldwide watersheds (Oudin. These data are regional averages. The mid-point of the region is +0. The latitude and the longitude values indicate the lower left corner of a 1x1 degree region. If the user introduces decimals to the latitude/longitude values. Boundaries of the -90/-180 region are -90 to -89 (south) and -180 to -179 (west). not point data. constant value of 2. If the user introduces decimals to the latitude/longitude values. The last region. Negative values are south and west.gov/sse). Re : extra-terrestrial radiation [MJ m-2 d-1] . positive values are north and east. λ: latent heat of vaporization.5 added to the latitude/longitude value. is bounded by 89 to 90 (north) and 179 to 180 (east). λ: latent heat of vaporization. ρ: water density.5 added to the latitude/longitude value. c) Oudin Oudin (2004) proposes following equation for the calculation of ETP: ETP  CoeffETP  ETP  0 Re T  5  if T  5    100 if T  5 (A. Ta : air temperature [C].11) if T  5 with ETP: potential evapotranspiration [m/d]. sponsored by the NASA’s Applied Science Program (http://eosweb. constant value of 1’000 [kg/m3] . 89/180.June 2005). T: Air temperature [°C] . Negative values are south and west. Latitude are only necessary for obtaining Re values The extra-terrestrial radiation Re is calculated as follows: RS MINERVE – Technical Manual Page 12/94 . ρ: water density. 2004). 89/180. Rg: global radiation [MJ/m2/day]. b) McGuinness McGuinness et Bordne (1972) proposes next ETP calculation: ETP  CoeffETP  Rg Ta  5     68 ETP  0 if T  5 (A. the RS MINERVE program calculates the nearest integer value for Rg calculations. Oudin method coefficients (5 and 100) were optimized for the hydrological modelling. not point data. positive values are north and east. constant value of 2. Rg data takes into account 22 year monthly average (July 1983 . The global radiation Rg is obtained in [kWh/m2/day] from the Global horizontal radiation dataset provided by the Surface meteorological and Solar Energy (SSE) web portal. The mid-point of the region is +0. constant value of 1’000 [kg/m3] .12) with ETP: potential evapotranspiration [m/d].Chapter 2: Hydrological models description Rg data takes into account 22 year monthly average (July 1983 . These data are regional averages.larc.26 [MJ/kg]. the RS MINERVE program calculates the nearest integer value for Rg calculations. The last region. This data comes as a grid (latitude and longitude) and is composed of monthly averaged values.nasa. Rg value is a location dependent (latitude and longitude) monthly average of the global radiation.June 2005). The latitude and the longitude values indicate the lower left corner of a 1x1 degree region. is bounded by 89 to 90 (north) and 179 to 180 (east).26 [MJ/kg]. Boundaries of the -90/-180 region are -90 to -89 (south) and -180 to -179 (west).

Dm :day of the month [-]. d) Uniform ETP The user can also set a uniform ETP for the whole simulation period and for the entire basin.  : latitude.14)   arccos(  tan( )  tan( ))   0.  : hour angle of the sun [rad]. negative in the south hemisphere [rad]. 1 to 12 [-] .16) if month  3 if month  3 and leap year  true (A. month : month of the year. Jd : Julian day [-] .409  sin( (A. ETP  CoeffETP  X (A.17) if month  3 and leap year  false with dr: relative distance Sun-Earth [-] .15) 2   J  1.Chapter 2: Hydrological models description Re  37. RS MINERVE – Technical Manual Page 13/94 .13) (A.033  cos( 2    Jd ) 365 (A.  : solar declination [rad] .6  dr  (  sin( )  sin( )  sin()  cos( )  cos( )) dr  1  0.39) 365 month  30  Dm 9 month J d  275   31  Dm 9 month J d  275   32  Dm 9 J d  275  (A.18) with X: uniform ETP [mm/d].

1 Melt coefficient due to liquid precipitation 0.2)  0 if T  Tcp1   (T  Tcp1 ) /(Tcp 2  Tcp1 ) if Tcp1  T  Tcp 2  1 if T  Tcp 2 RS MINERVE – Technical Manual (B.Chapter 2: Hydrological models description 2.1 to B.0125 Minimum critical temperature for liquid 0 precipitation Maximum critical temperature for solid 4 precipitation Critical snowmelt temperature 0 Initial snow height Initial relative water content in the snow pack - In a first step.5 to 20 Critical relative water content of the snow pack 0. the precipitation is divided into a solid precipitation (N) and into a liquid precipitation (P*) as a function of the temperature (equations B.3 Snow-GSM model description The Snow-GSM model (Figure 2) is composed of two sub-models which simulate the transient evolution of the snow pack (accumulation and melt) as a function of the temperature (T) and precipitation (P) producing an equivalent precipitation (Peq) which can be used as an input variable by the infiltration or the Glacier-GSM model.3): P*    P (B. Figure 2 Snow-GSM model Table 2 List of parameters and initial conditions for the Snow-GSM model Object Name An ThetaCri bp Tcp1 Units mm/°C/day d/mm °C Tcp2 °C Tcf HIni ThetaIni °C m - SnowGSM Decription Regular range Degree-day snowmelt coefficient 0.3) Page 14/94 .1) N  (1   )  P (B.

If the observed temperature is lower than Tcp1 only solid precipitation is produced. Tcf = 0 °C) but can be also be calibrated for some cases. T: temperature [°C]. θcr. Tcp1 = 0 °C.8):   WN / H (B. Tcp1: minimum critical temperature for liquid precipitation [°C].1. An: degree-day snowmelt coefficient [L/T/°C].  : separation factor.4) dH / dt  N  M N M N  N  H / dt (B. The parameters bp. dt: time step [T]. P: precipitation [L/T]. RS MINERVE – Technical Manual Page 15/94 . The variables for the initial situation associated to this model are θ (Theta in RS MINERVE) and H. WN: water content [L]. varying its content as a function of melt or freezing.6 to B.Chapter 2: Hydrological models description with P*: liquid precipitation [L/T]. Tcp2 and Tcf can be assumed as constant (bp = 0.0125. Tcp1. bp: precipitation coefficient due to melt [T/L].7) (B. If the temperature is higher than Tcp2 only liquid precipitation is produced. If the temperature observed is found between these two critical values liquid and solid precipitation are produced. θcr = 0. H: snow height [L]. Tcf: critical snowmelt temperature [°C]. The input variables of the model are precipitation (P) and temperature (T). Peq: equivalent precipitation [L/T]. Tcp2: maximum critical temperature for solid precipitation [°C]. N: solid precipitation [L/T]. The snowmelt calculation is performed as follows: M N  An  (1  bp  P*)  (T  Tcf ) if T  Tcf M N  An  (T  Tcf ) if T  Tcf (B. The solid precipitation (N) is used as input for the snow pack.5) M N  WN / dt with MN: snowmelt or freezing [L/T]. The equivalent precipitation (Peq) is produced by the water content of the snow (equations B. The parameters to adjust are An and θcr. Tcp2 = 4 °C. the output value is the equivalent precipitation (Peq). θcr (ThetaCri in RS MINERVE): critical relative water content in the snow pack [-].8) with θ (Theta in RS MINERVE): relative water content in the snow pack [-].6) Peq  P * WN / dt if H  0 Peq  0 if H  0 et    cr Peq  (  cr )  H / dt if H  0 et    cr dWN / dt  P *  M N  Peq (B.

. 2005).1 to 5 0.4 Glacier-GSM model description In the Glacier-GSM model (Schäfli & al.5 to 20 0.Chapter 2: Hydrological models description 2. The total discharge of the glacier also depends on the transfer processes within the linear snow and glacier reservoirs R N and RGL. the glacier melt (Figure 3) depends on the temperature and the presence of snow on the glacier.1 to 5 0 - Page 16/94 . Figure 3 Glacier-GSM model Table 3 List of parameters and initial conditions for the Glacier-GSM model Object GlacierGSM Name A Agl Kgl Kgl Units m2 mm/°C/day 1/d 1/d Tcg QSnowIni QGlacierIni °C m3/s m3/s RS MINERVE – Technical Manual Description Surface of glacier Degree-day icemelt coefficient Release coefficient of icemelt reservoir Release coefficient of snowmelt reservoir Critical glacier melt temperature Initial outflow of linear snow reservoir Initial outflow of linear glacier reservoir Regular Range >0 0.

2.5) with PeqGL: glacier melt [L/T]. the temperature (T) and the equivalent precipitation (Peq). RS MINERVE – Technical Manual Page 17/94 .4) Qglacier  K gl  H GL  A (C. Tcg (which can be assumed as constant. And the total discharge is the addition of Qglacier and Qsnow: Q  Qsnow  Qglacier (C.Chapter 2: Hydrological models description A gross precipitation (or equivalent precipitation Peq from the Snow-GSM model) is transferred to the linear snow reservoir (Rsn). The variables associated to this model are Hsnow. Tcg: critical glacier melt temperature [°C]. The water intensity produced by the glacier melt (PeqGL) is transferred to the linear glacier reservoir (Rgl) and the resulting discharge (Qglacier) at the outlet of the sub-catchment. ]. A: glacier surface [L2].2) with Peq: equivalent precipitation [L/T]. θ. Qglacier: outflow of linear glacier reservoir [L3/T].1) Qsnow  K sn  H Rsn  A (C. Agl: degree-day glacier melt coefficient [L/T/ºC]. Kgl: coefficient of linear glacier reservoir [1/T]. Kgl and Ksn.5: PeqGL  0 if T  Tcg or H snow  0 PeqGL  Agl  (T  Tcr ) if T  Tcg and H snow  0 (C. HRgl: level of glacier melt reservoir [L]. The parameters to adjust are Agl. The parameter A is supposed to be constant. The output is the total discharge (Q) at the model outlet. HRsn: level in linear snow reservoir [L]. generating an outflow according to equations C. The model inputs are the snow level (Hsnow). Qsnow: outflow of linear snow reservoir [L3/T]. HRsn.6) with Q: total outflow [L3/T]. and HRgl.1 and C.3) dH Rgl / dt  PeqGL  K gl  H Rgl (C. Tcg=0°C). The glacier melt sub-model only provides a discharge when the snow level is zero (Hsn=0). dH Rsn / dt  Peq  K N  H Rsn (C.3 to C. Ksn: release coefficient of linear snow reservoir [1/T]. The glacier melt Qglacier is defined according to equations C.

2) (D. The parameter A is supposed to be constant.. HMax: capacity of infiltration reservoir [L]. RS MINERVE – Technical Manual Page 18/94 . Figure 4 GR3 Infiltration model Table 4 List of parameters and initial conditions for the GR3 model Object GR3 Name A HMax K HIni Units m2 m 1/s m Description Surface of infiltration Maximum height of infiltration reservoir Release coefficient of infiltration reservoir Initial level in infiltration reservoir Regular Range >0 0 to 2 0.4) (D. The variable for the initial state associated to this model is H. 1988.5 Infiltration (GR3) model description The gross intensity (or equivalent precipitation Peq coming from the Snow-GSM model) and potential evapotranspiration (ETP) are introduced in the GR3 model (Michel and Edijatno. K: release coefficient of infiltration reservoir [1/T]. Q: base discharge [L3/T]. H: level in infiltration reservoir [L].Chapter 2: Hydrological models description 2. iNet: net intensity [L/T]. 1998). ETR: real evapotranspiration [L/T]. The parameters to adjust are K and HMax.1) is computed as follows: iInf  Peq  (1  ( H / H Max ) 2 ) if H  H Max iInf  0 if H  H Max ETR  ETP  H / H Max if H  H Max ETR  ETP if H  H Max iNet  Peq  iInf (D.1 - The infiltration reservoir (Figure D.1) (D. A: infiltration surface [L2].3) Q  K H  A if H  H Max Q  K  H Max  A if H  H Max dH / dt  iInf  ETR  Q / A (D. Peq: equivalent precipitation [L/T].5) with ETP: potential evapotranspiration [L/T]. iInf: infiltration intensity [L/T]. Consuegra & al.00025 to 0.

P eq. RS MINERVE – Technical Manual Page 19/94 . the base discharge (Q) and the real evapotranspiration (ETR) are obtained. the net intensity (i Net). and the potential evapotranspiration (ETP). As a result.Chapter 2: Hydrological models description The input variables of the model are the equivalent precipitation (or gross intensity).

6 Runoff (SWMM) model description The SWMM (Storm Water Management Model) model presented hereafter was developed by Metcalf and Eddy (1971). supplied by a hyetograph of net rainfall (iNet). A: run-off surface [L2]. A) are supposed to be constant. Jo: average slope of the plane [-]. L.1) ir  K  J o  H 5 / 3 / L (E. The other parameters (Jo. L: length of the plane [L]. Figure 5 SWMM Runoff model Table 5 List of parameters and initial conditions for the SWMM model Object SWMM Name A L J0 K HIni Units m2 m 1/3 m /s m Description Regular Range Surface of runoff >0 Length of the plane >0 Runoff slope >0 Strickler coefficient 0. provides a hydrograph downstream of the surface (Q). ir: runoff intensity [L/T]. RS MINERVE – Technical Manual Page 20/94 . The parameter to adjust is K. iNet: net intensity [L/T]. The SWMM model.2) Q  ir  A (E.3) with H: runoff water level downstream of the surface [L]. The variable for the initial condition associated to the model is Hr.3: dH / dt  2  (i Net  ir ) Hr  0 (E.1 to E.Chapter 2: Hydrological models description 2.1 to 90 Initial water level downstream of the surface - The transfer of the net intensity to an impermeable surface is carried out by the help of a non-linear transfer reservoir Figure 5) depending on the equations E. K: Strickler coefficient [L1/3/T].

the sub-model of the glacier melt creates a flow when the height of snow is zero (Hsnow=0). The final flow (Qtot) produced by the sub-catchment is the addition of the two flows (Qglacier and Qsnow). Besides. the snow model creates an equivalent precipitation (Peq) which is transferred to the glacier model. two corresponding to the SnowGSM model and the other three corresponding to the Glacier-GSM model. Figure 6 GSM model RS MINERVE – Technical Manual Page 21/94 . This glacier flow (PeqGL) is transferred to the linear glacier reservoir (Rgl) and the resulting flow (Qglacier) to the outlet of the sub-catchment. The same accounts for the height of the snow (Hsnow) and the temperature (T). From the inputs of precipitation (P) and temperature (T). The hereafter present model allows an easy construction of this kind of composition.7 GSM model description The GSM model (Figure 6) is composed of 5 sub-models. In the glacier model the equivalent precipitation is transferred to the linear snow reservoir (Rsn) and finally to the outlet of the sub-catchment (Qsnow).Chapter 2: Hydrological models description 2.

the precipitation is divided into a solid precipitation (N) and into a liquid precipitation (P*) as a function of the temperature (equations F.5 to 20 ThetaCri Critical relative water content of the snow pack 0. The equivalent precipitation (Peq) is produced by the water content of the snow (equations F. WN: water content [L]. The snowmelt calculation is performed as follows: M N  An  (1  bp  P*)  (T  Tcf ) if T  Tcf M N  An  (T  Tcf ) if T  Tcf (F.5 to 20 Kgl 1/d Release coefficient of icemelt reservoir 0. The solid precipitation (N) is used as input for the snow pack. If the temperature is higher than Tcp2 only liquid precipitation is produced. varying its content as a function of melt or freezing.1 to 5 Tcp1 °C Minimum critical temperature for liquid precipitation 0 Tcp2 °C Maximum critical temperature for solid precipitation 4 Tcf °C Critical snowmelt temperature 0 Tcg °C Critical glacier melt temperature 0 HsnowIni m Initial snow height 3 QsnowIni m /s Initial outflow of linear snow reservoir QglacierIni m3/s Initial outflow of linear glacier reservoir ThetaIni Initial relative water content in the snow pack - In a first step. An: degree-day snowmelt coefficient [L/T/°C]. Tcf: critical snowmelt temperature [°C]. Tcp1: minimum critical temperature for liquid precipitation [°C].3) with P*: liquid precipitation [L/T]. P: precipitation [L/T].4) dH / dt  N  M N M N  N  H / dt (F. N: solid precipitation [L/T].2)  0 if T  Tcp1   (T  Tcp1 ) /(Tcp 2  Tcp1 ) if Tcp1  T  Tcp 2  1 if T  Tcp 2 (F. H: snow height [L].  : separation factor.3): P*    P (F. If the observed temperature is lower than Tcp1 only solid precipitation is produced.1) N  (1   )  P (F.0125 Agl mm/°C/day Degree-day icemelt coefficient 0.1 to F. dt: time step [T].1 to 5 Ksn 1/d Release coefficient of snowmelt reservoir 0.1 bp d/mm Melt coefficient due to liquid precipitation 0. If the temperature observed is found between these two critical values liquid and solid precipitation are produced.5) M N  WN / dt with MN: snowmelt or freezing [L/T].8): RS MINERVE – Technical Manual Page 22/94 .Chapter 2: Hydrological models description Table 6 List of parameters and initial conditions for the GSM model Object GSM Name Units Description Regular Range A m2 Surface of infiltration >0 An mm/°C/day Degree-day snowmelt coefficient 0. T: temperature [°C].6 to F. Tcp2: maximum critical temperature for solid precipitation [°C]. bp: precipitation coefficient due to melt [T/L].

13) with PeqGL: glacier melt [L/T]. Tcp1. Hsnow.1. The parameters to adjust are An. Peq: equivalent precipitation [L/T].9 and F. A: glacier surface [L2]. HRgl: level of glacier melt reservoir [L].Tcf and Tcg can be assumed as constant (bp = 0. θcr = 0. HRsn: level in linear snow reservoir [L]. And the total discharge is the addition of Qglacier and Qsnow: Q  Qsnow  Qglacier (F.7) (F.Chapter 2: Hydrological models description   WN / H (F. and HRgl.0125. Qsnow: outflow of linear snow reservoir [L3/T]. The equivalent precipitation Peq is then transferred to the linear snow reservoir (Rsn). RS MINERVE – Technical Manual Page 23/94 . The variables for the initial situation associated to this model are θ (Theta in RS MINERVE). Kgl: coefficient of linear glacier reservoir [1/T]. ]. Ksn: release coefficient of linear snow reservoir [1/T]. The parameter A is supposed to be constant. The input variables of the model are the precipitation (P) and the temperature (T). Tcf = 0 °C. The output is the total discharge (Qtot) at the model outlet.8) with θ (Theta in RS MINERVE): relative water content in the snow pack [-]. θcr (ThetaCri in RS MINERVE): critical relative water content in the snow pack [-]. θcr. Tcg = 0 °C). The water intensity produced by the glacier melt (PeqGL) is transferred to the linear glacier reservoir (Rgl) and the resulting discharge (Qglacier) at the outlet of the sub-catchment.10) with Peq: equivalent precipitation [L/T]. Tcp1 = 0 °C.12) Qglacier  K gl  H Rgl  A (F. Qglacier: outflow of linear glacier reservoir [L3/T].9) (F. The glacier melt sub-model only provides a discharge when the snow level is zero (Hsn=0). Tcp2. Tcp2 = 4 °C. generating an outflow according to equations F. The parameters bp.6) Peq  P * WN / dt if H  0 Peq  0 if H  0 et    cr Peq  (  cr )  H / dt if H  0 et    cr dWN / dt  P *  M N  Peq (F. HRsn.14) with Q: total outflow [L3/T].10. Agl.11 to F. Agl: degree-day glacier melt coefficient [L/T/ºC]. dH Rsn / dt  Peq  K N  H Rsn Qsnow  K sn  H Rsn  A (F. Tcg: critical glacier melt temperature [°C].13: PeqGL  0 if T  Tcg or H snow  0 PeqGL  Agl  (T  Tcr ) if T  Tcg and H snow  0 (F.11) dH Rgl / dt  PeqGL  K gl  H Rgl (F. The glacier melt Qglacier is defined according to equations F. Kgl and Ksn.

Figure 7 SOCONT Model RS MINERVE – Technical Manual Page 24/94 . the Snow-GSM model simulates the transient evolution of the snow pack (melt and accumulation) as a function of the temperature (T) and the precipitation (P). thus providing an equivalent precipitation (P eq) that is used as input by the GR3 model.Chapter 2: Hydrological models description 2.8 SOCONT model description In the SOCONT model (Figure 7). The GR3 model also takes into account the potential evapotranspiration (ETP) and provides the net intensity to the SWMM model.

1 to G. P: precipitation [L/T]. If the temperature observed is found between these two critical values liquid and solid precipitation are produced.1 Length of the plane >0 Runoff slope >0 Strickler coefficient 0. N: solid precipitation [L/T]. Tcp1: minimum critical temperature for liquid precipitation [°C].2)  0 if T  Tcp1   (T  Tcp1 ) /(Tcp 2  Tcp1 ) if Tcp1  T  Tcp 2  1 if T  Tcp 2 (G. The solid precipitation (N) is used as input for the snow pack.0125 Minimum critical temperature for liquid 0 precipitation Maximum critical temperature for solid 4 precipitation Critical snowmelt temperature 0 Maximum height of infiltration reservoir 0 to 2 Release coefficient of infiltration reservoir 0. the precipitation is divided into a solid precipitation (N) and into a liquid precipitation (P*) as a function of the temperature (equations G.5) M N  WN / dt RS MINERVE – Technical Manual Page 25/94 .1 Melt coefficient due to liquid precipitation 0.4) dH / dt  N  M N M N  N  H / dt (G.3) with P*: liquid precipitation [L/T].1) N  (1   )  P (G. The snowmelt calculation is performed as follows: M N  An  (1  bp  P*)  (T  Tcf ) if T  Tcf M N  An  (T  Tcf ) if T  Tcf (G. Tcp2: maximum critical temperature for solid precipitation [°C].Chapter 2: Hydrological models description Table 7 List of parameters and initial conditions for the SOCONT model Object Name A An ThetaCri bp Tcp1 Units m2 mm/°C/day d/mm °C Tcp2 °C Tcf SOCONT HGR3Max KGR3 L J0 Kr HsnowIni HGR3Ini HrIni °C m 1/s m 1/3 m /s m m m ThetaIni - Description Regular Range Surface of infiltration >0 Degree-day snowmelt coefficient 0.1 to 90 Initial snow height Initial level in infiltration reservoir Initial runoff water level downstream of the surface Initial relative water content in the snow pack - In a first step.00025 to 0.3): P*    P (G. T: temperature [°C].  : separation factor.5 to 20 Critical relative water content of the snow pack 0. varying its content as a function of melt or freezing. If the temperature is higher than Tcp2 only liquid precipitation is produced. If the observed temperature is lower than Tcp1 only solid precipitation is produced.

the infiltration reservoir is computed as follows: iInf  Peq  (1  ( H / H Max ) 2 ) if H  H Max iInf  0 if H  H Max ETR  ETP  H / H Max if H  H Max ETR  ETP if H  H Max iNet  Peq  iInf (G. dt: time step [T]. WN: water content [L]. K: Strickler coefficient [L1/3/T]. bp. ir: runoff intensity [L/T].7) (G.14 to G. θcr (ThetaCri in RS MINERVE): critical relative water content in the snow pack [-]. HGR3 and Hr. The variables for the initial situation associated to this model are θ (Theta in RS MINERVE). Peq: equivalent precipitation [L/T]. A: infiltration surface [L2]. Then. The RS MINERVE – Technical Manual Page 26/94 . iNet: net intensity [L/T]. HGR3Max and Kr. Peq: equivalent precipitation [L/T].8) with θ (Theta in RS MINERVE): relative water content in the snow pack [-]. H: level in infiltration reservoir [L].10) (G. iInf: infiltration intensity [L/T]. An: degree-day snowmelt coefficient [L/T/°C].14) ir  K  J o  H 5 / 3  L (G. Jo: average slope of the plane [-].9) (G.13) with ETP: potential evapotranspiration [L/T]. bp: precipitation coefficient due to melt [T/L]. The equivalent precipitation (Peq) is produced by the water content of the snow (equations G. θcr. Tcf: critical snowmelt temperature [°C]. The transfer of the net intensity to an impermeable surface is carried out by the help of a non-linear transfer reservoir depending on the equations G.15) Q  ir  A (G.8):   WN / H (G.Chapter 2: Hydrological models description with MN: snowmelt or freezing [L/T]. ETR: real evapotranspiration [L/T]. L: length of the plane [L]. iNet: net intensity [L/T].12) (G. H: snow height [L].6 to G. K: release coefficient of infiltration reservoir [1/T].11) Q  K H  A if H  H Max Q  K  H Max  A if H  H Max dH / dt  iInf  ETR  Q / A (G. KGR3.16) with H: runoff water level downstream of the surface [L]. HMax: capacity of infiltration reservoir [L]. Hsnow. Q: base discharge [L3/T]. The parameters to adjust are An. A: run-off surface [L2].6) Peq  P * WN / dt if H  0 Peq  0 if H  0 et    cr Peq  (  cr )  H / dt if H  0 et    cr dWN / dt  P *  M N  Peq (G.16: dH / dt  2  (i Net  ir ) Hr  0 (G.

The input variables of the model are the precipitation (P). the temperature (T) and the potential evapotranspiration (ETP). L. The other parameters (Jo. Tcf = 0 °C). Tcp2 and Tcf can be assumed as constant (bp = 0. Tcp1 = 0 °C. θcr = 0. The output value is the equivalent precipitation (P eq). composed by the base discharge (QGR3) and the run-off discharge (Qr). is finally transferred to the outlet of the sub-catchment. Tcp1. A) are supposed to be constant.Chapter 2: Hydrological models description parameters bp.1. Tcp2 = 4 °C. RS MINERVE – Technical Manual Page 27/94 .0125. θcr. The outflow discharge Qtot.

The structure of the implemented model is presented in the Figure 8.9 HBV model description The integrated rainfall-runoff model HBV (Bergström. Figure 8 HBV model RS MINERVE – Technical Manual Page 28/94 . a humidity reservoir and two (upper and lower) soil storage reservoirs. 1976. 1992) is composed of a snow function.Chapter 2: Hydrological models description 2.

050 to 0.3).8 - The precipitation is first divided into snowfall (SF) and rainfall (RF) as a function of the temperature (equations H. T: temperature [°C]. The snowmelt calculation is performed as follows: M sn  CFMax  (T  TTSM ) if T  TTSM M sn  CFR  CFMax  (T  TTSM ) if T  TTSM (H. α: separation factor. TTInt: temperature interval for rain/snow mixing [°C]. only solid snowfall is produced. If the observed temperature is lower than TT-TTInt/2. The snowfall (SF) is used as input for the snow pack.4 0 to 0. P: precipitation [L/T]. If the observed temperature is found between these values.5 to 20 0. varying its content as a function of melt or freezing.5 0.15 0 to 0. If the temperature is higher than TT+TTInt/2.05 0. only rainfall is produced.3) if T  TT  TTInt / 2 with RF: rainfall [L/T]. both rainfall and snowfall are produced.5) M sn   H water / dt RS MINERVE – Technical Manual Page 29/94 . RF    P (H.1 0 to 3 0 to 3 0 1 to 5 0.10 0.1) SF  (1   )  P (H.05 to 0.65 0 to 0. TT: threshold temperature for rain/snow [°C].01 to 0.Chapter 2: Hydrological models description Table 8 List of parameters and initial conditions for the HBV model Object Name Units Description A m2 Surface of the basin CFMax mm/°C/day Melting factor CFR Refreezing factor CWH Critical relative water content of the snow pack TT °C Threshold temperature of rain/snow TTInt °C Temperature interval for rain/snow mixing TTSM °C Threshold temperature for snow melt Beta Model parameter (shape coefficient) FC m Maximum soil storage capacity PWP m Soil permanent wilting point SUMax m Upper reservoir water level threshold Kr 1/d Near surface flow storage coefficient Ku 1/d Interflow storage coefficient Kl 1/d Baseflow storage coefficient Kperc 1/d Percolation storage coefficient HsnowIni m Initial snow height WHIni Initial relative water content in the snow pack HIni m Initial humidity SUIni m Initial upper reservoir water level SLIni m Initial lower reservoir water level HBV Regular Range >0 0.4) dH snow / dt  SF  M sn M sn  SF  H snow / dt (H.2)  0 T  (TT  TTInt / 2) TTInt  1  if T  TT  TTInt / 2 if TT  TTInt / 2  T  TT  TTInt / 2 (H.65 0. SF: snowfall [L/T].030 to 0.1 to H.

9.13) (H. Hum: Humidity [L].10) if Hum  PWP dHum / dt  ( Peq  iRecharge )  ETR Hum  0 (H.14) Page 30/94 .6 to H. near surface (or run-off) flow is calculated depending on the water level in the Upper reservoir (SU) and its threshold.Chapter 2: Hydrological models description with Msn: snowmelt or freezing [L/T]. corresponding to the upper soil storage and producing the interflow.8): WH  H water / H snow (H. Hsnow: snow height [L]. Hwater: water content [L]. CFMax: degree-day melting factor [L/T/°C]. Finally.6) Peq  RF  H water / dt if H snow  0 Peq  0 if H snow  0 et WH  CWH Peq  (WH  CWH )  H snow / dt if H snow  0 et WH  CWH dH water / dt  RF  M sn  Peq (H. as presented in equation H. Qr  K r  ( SU  SU Max )  A if SU  SU Max Qr  0 if SU  SU Max (H. The Upper reservoir (or interflow reservoir). dt: time step [T]. CWH: critical relative water content in the snow pack [-]. The equivalent precipitation (Peq) is produced by the water content of the snow (equations H. CFR: refreezing factor [-]. P eq) and outputs (Recharge intensity and ETR) as presented in equation H.11) with iRecharge: Reservoirs recharge intensity [L/T] .8) with WH: relative water content in the snow pack [-]. The calculation of the recharge is carried out depending on a model parameter Beta. Kr: Near surface flow storage coefficient [1/T].12) with Qr: Near surface flow (or run-off flow) [L3/T]. ETR: Evapotranspiration [L/T] .11:  Hum  i Recharge     FC  ETR  ETP  Beta Hum PWP ETR  ETP  Peq (H.9) if Hum  PWP (H. as well as on a flow storage coefficient Kr.7) (H. SU: Upper reservoir water level [L]. is computed as follows: dSU / dt  iRecharge  K perc  Ku  SU  Qr / A SU  0 i Perc  K perc  SU RS MINERVE – Technical Manual (H. SUMax: Upper reservoir water level threshold [L]. Peq: Equivalent precipitation [L/T]. A: Basin surface [L2]. FC: Maximum soil storage capacity [L] . Beta: Model parameter (shape coefficient) [-]. the humidity of the soil (Hum) is performed taking into account the input (Equivalent precipitation. ETR is calculated as shown in equation H. Peq: equivalent precipitation [L/T]. Then. ETP: Potential evapotranspiration [L/T] . TTSM: critical snowmelt temperature [°C].10. PWP: Soil permanent wilting point [L].

HumIni. Beta.Chapter 2: Hydrological models description Qu  Ku  SU  A (H.15) with Kperc: Percolation storage coefficient [1/T] . The initial conditions associated to this model are HsnowIni. WHIni. PWP. Kl: Baseflow storage coefficient [1/T]. the temperature (T) and the potential evapotranspiration (ETP). is calculated as presented in equations H. Ql: Baseflow [L3/T]. CWH. iPerc: Percolation intensity [L/T] . TTSM. Qu: Interflow [L3/T]. The parameters to adjust are CFMax.18) with Qtot: Total outflow [L3/T]. Kr. The model inputs are the precipitation (P). FC. The parameter A is supposed to be constant.16) SL  0 Ql  Kl  SL  A (H. the lower reservoir (or baseflow reservoir). Ku: Interflow storage coefficient [1/T] . Kl and Kperc. SUIni and SLIni.17 dSL / dt  iPerc  Kl  SL (H.16 and H. CFR. TTInt. SUmax. And finally the total outflow is: Qtot  Qr  Qu  Ql (H.17) with SL: Lower reservoir water level [L]. corresponding to the lower soil storage. The output is the total discharge at the model outlet (Qtot). Ku. Afterwards. RS MINERVE – Technical Manual Page 31/94 . TT.

10 GR4J model description The GR4J model is a global hydrological model with four parameters developed by Perrin et al. It is an empirical model (Figure 9). Unit hydrographs are also associated for the hydrological behaviour of the basin. (2003). Figure 9 GR4J model RS MINERVE – Technical Manual Page 32/94 .Chapter 2: Hydrological models description 2. It takes into account the humidity and contains two reservoirs (production and routing). but its structure is similar to the conceptual models.

Es : evapotranspiration from the production reservoir [L/T].5 : RS MINERVE – Technical Manual Page 33/94 . Similarly.2. The percolation (iPerc) flowing from the production reservoir is then calculated as follows: 1  4 4      4 S  Ps  Es   dt     d   iPerc  S  Ps  Es   dt   1  1    dt X1    9      (I.9 Initial water content in the production reservoir Initial water level in the routing reservoir - First calculation is the neutralization of Peq by ETP for determining a net intensity (Pn) and a net evapotranspiration (En) according to equations I. Finally.1 to 1.3) with Ps: rain intensity feeding the production reservoir [L/T].1 to 2. X1 : maximum capacity of production reservoir S [L] .003 Capacity of routing store 0.2.02 to 0.Chapter 2: Hydrological models description Table 9 List of parameters and initial conditions for the GR4J model Object GR4J Name A X1 X2 X3 X4 SIni RIni Units m2 m m m d m m Description Regular Range Surface of the basin >0 Capacity of production store 0.3.   S 2   Pn  dt   X 1  1      tanh   X X 1   d   1  Ps  dt  S   Pn  dt   1     tanh  X1   X1  (I.2)   S   En  dt     tanh  S   2    X X d 1  1     Es  dt   En  dt  S    tanh  1  1  X1    X1  (I. ETP : potential evapotranspiration [L/T] .3 UH time base 1. En : net evapotranspiration [L/T].005 to 0.2 Water exchange coefficient -0. Pn  Peq  ETP En  0 Pn  0 En  ETP  Peq if P  ETP (I.1 and I. a part (Ps) of Pn feed the production reservoir (S) as presented in equation I. Peq : Equivalent precipitation [L/T] .1) if P  ETP with Pn : net rain intensity [L/T] .4) with iPerc : percolation intensity [L/T]. the variation in the production reservoir S is obtained according to equation I. S : water content in the production reservoir [L]. when En is not zero. When Pn is not zero. an evapotranspiration (Es) from the production reservoir is calculated depending on the water level in the production reservoir as described in equation I.

l = integer value of X4 + 1 [T] .8) 5 1  t 2  SH 2t  1    2  2  X 4  SH 2t  1 if X 4  t  2  X 4 if t  2  X 4 UH1 j  SH1 j  SH1 j 1 UH 2 j  SH 2 j  SH 2 j 1 (I. SH2 : hydrograph 2 from S [-].6) with iPr : intensity flowing to the routing reservoir [L/T]. UH2 : hydrograph from SH2 [-]. [L/T].7 and I.11 and I. M j : day (integer) [T].9   UH1 j  i Prk  j l  (I.13 (physical interpretation of this function is not direct). UH1 and UH2 depend on parameter X4. SH 1t  0 if t  0  t   SH 1t    X4  SH 1t  1 5 2 if 0  t  X 4 if t  X 4 SH 2t  0 SH 2t  (I. Q1: Unitary flow from the UH2 hydrograph [L/T].14.Chapter 2: Hydrological models description dS / dt  Ps  Es  iPerc S 0 (I. UH1 : hydrograph from SH1 [-].12.1   UH 2 j  i Prk  j l  (I.5) The quantity of water (iPr) which goes to the routing reservoir is provided by : iPr  iPerc  Pn  Ps  (I. SH1 and SH2 are defined depending on the time step t as presented in equation I. t: time step [T]. Afterwards. Q9  0. iPr is divided in two flow components: 90% are routed by a unit hydrograph UH1 and a routing reservoir and 10% by a symmetrical unit hydrograph UH2. RS MINERVE – Technical Manual Page 34/94 .12) l j 1 m j 1 with Q9: Unitary flow from the UH1 hydrograph. for the calculation of unit hydrographs. which is the base time. Then.7) if t  0 1  t    2  X 4  5 2 if 0  t  X 4 (I.11) Q1  0. m = integer value of 2*X4+1 [T]. the routing reservoir (R) is emptied by a routing discharge (Qr’) as described in equation I. at each time step k.8. X4: base time [T]. Then. The ordinates of the hydrographs are calculated from curves SH1 and SH2.9) (I. which correspond to the cumulated hydrograph. Then. an exchange function (iF) with the non-atmospheric outside produces an underground water exchange according to I.10) with SH1 : hydrograph 1 from S [-]. outflows Q9 and Q1 from the two hydrographs correspond to the convolution of previous intensities according to the repartition provided by the discretized hydrograph and calculated as presented in equations I. The ordinates of UH1 and UH2 are then calculated.

as presented in equation I.16) with Qd’: unitary complementary outflow [L/T]. X2 : water exchange coefficient. A: Basin surface [L2]. positive for contributions. R : water level in the routing reservoir [L] . The variables associated to this model are SIni and RIni. Qtot. Qd: complementary outflow [L3/T].13) 1  4 4      R  Q9  iF   dt    d   Qr'  R  Q9  iF   dt   1  1   dt X3         (I.18) Qtot  Qr  Qd (I. Finally. X3 and X4. RS MINERVE – Technical Manual Page 35/94 . The variation of the water level in the routing reservoir is: dR / dt  Q9  iF  Qr R0 (I. The parameters to adjust are X1.19) with Qr: outflow from the routing reservoir. The output is the total discharge at the model outlet (Qtot).17 and I. The model inputs are the equivalent precipitation (Peq) and the potential evapotranspiration (ETP).14) with iF : exchange function [L/T] . X2. The parameter A is supposed to be constant after its calculation. the model outflows are calculated as presented in equations I.18 and the total flow at the outlet. negative for losses due to infiltration or zero when no exchange is produced [L]. X3 : one day capacity of routing reservoir [L]. Qtot : total outflow [L3/T].19: Qr  Qr'  A (I. Qr’: unitary routing reservoir outflow [L/T].Chapter 2: Hydrological models description 7  R 2 d  iF  X 2   dt  X3  (I.17) Qd  Qd '  A (I.15) The outflow Q1 from hydrograph UH2 has also the same exchange for providing the complementary flow Qd’: Qd '  Q1  iF (I.

This model calculates the total discharge from the precipitation (P) and the potential evapotranspiration (ETP) depending on the parameters and initial conditions presented in Table 10. 1973.Chapter 2: Hydrological models description 2.11 SAC-SMA model description The SAC-SMA (Figure 10) or SACRAMENTO hydrological model was developed in the 70’s (Burnash et al.. for an efficient simulation of discharges. 1995) to optimize humidity characteristics into the soil. distributed into different level. Figure 10 SAC-SMA model RS MINERVE – Technical Manual Page 36/94 . with rational percolation characteristics. Burnash.

15 Upper Zone Free Water capacity 0. determined by:   The permanent impervious area fraction (Pctim). This area produces impervious and direct runoff QDirectRunoff from any rainfall. RS MINERVE – Technical Manual Page 37/94 . The basin may present permanent or temporarily impervious areas.3 Free Water storage Ratio of deep percolation from Lower Zone Free 0 to 0.03 Water storage Depletion rate of the Lower Zone supplemental 0.10 to 0.10 Interflow depletion rate from the Upper 0.03 to 0.005 to 0.80 The Lower Zone supplementary Free Water 0. It represents the maximum portion of reservoirs and temporary watersheds contained in the basin. The maximum fraction of an additional impervious area due to saturation (Adimp). A pervious portion of the soil mantle which distributes rainwater to the next storages.05 Riparian vegetarian area fraction 0 to 0.2 area due to saturation Permanent impervious area fraction 0 to 0.75 Zone Free Water storage Ratio of maximum and minimum percolation rates 10 to 350 Shape parameter of the percolation curve 1 to 4 Percolation fraction that goes directly to the 0 to0. marshes and other impervious material directly linked to the streamflow network.40 The Lower Zone primary Free Water capacity 0.02 to 0. the pervious area can produce runoff when rainfall rates are sufficiently heavy. lake surfaces. In addition.2 Upper Zone Tension Water capacity 0.40 capacity Fraction of Lower Zone Free Water not 0 to 1 transferable to Lower Zone Tension Water Depletion rate of the Lower Zone primary Free 0.Chapter 2: Hydrological models description Table 10 List of parameters and initial conditions for the SAC-SMA model Object SACSMA Name A Adimp Units m2 - Pctim Riva UztwMax UzfwMax Uzk m m 1/day Zperc Rexp Pfree - LztwMax LzfpMax LzfsMax m m m Rserv - Lzpk 1/day Lzsk 1/day Side - AdimIni UztwIni UzfwIni LztwIni LzfpIni LzfsIni m m m m m m Description Regular Range Surface of the basin >0 Maximum fraction of an additional impervious 0 to 0. two basic basin areas are taken into account:   An impervious portion of the soil mantle covered by streams.01 to 0.001 to 0.6 Lower Zone Free Water storages The Lower Zone Tension Water capacity 0.5 Water storages Initial Tension Water content of the Adimp area Initial Upper Zone Tension Water content Initial Upper Zone Free Water content Initial Lower Zone Tension Water content Initial Lower Zone Free supplementary content Initial Lower Zone Free primary content - When considering rainfall over a basin.01 to 0.05 to 0.

a transfer is produced from the Free Water storage to the Tension Water storage. The following concepts are defined for this reservoir:  The Upper Zone Tension Water Maximum capacity (UztwMax) is the maximum Tension Water storage admissible in the Upper Zone. where the input is the percolation from the Upper Zone.5: 𝑇= 𝑈𝑧𝑡𝑤𝐶 𝑈𝑧𝑡𝑤𝑀𝑎𝑥 (J. but also from transfers from the Upper Zone Free Water storage.  The Upper Zone Tension Water Content (UztwC) is the Tension Water stored in this Upper Zone at any given time. UztwMax: the Upper Zone Tension Water capacity RS MINERVE – Technical Manual Page 38/94 .Chapter 2: Hydrological models description Thus. the permanent pervious area fraction of the basin is calculated following equation J. If the relative content of Tension Water (ratio between its content and its maximum storage) is smaller than the relative content of Free Water. purely impervious or mixed. As shown in the Figure 10. F: relative content of Upper Zone Free Water [-]. A Lower Zone. It is worth mentioning that the behaviour of the additional impervious area might be purely pervious.4) if 𝑇 ≥ 𝐹 (J.2) 𝐹= 𝑈𝑧𝑓𝑤𝐶 𝑈𝑧𝑓𝑤𝑀𝑎𝑥 (J. The main components of the model and their flow transfers are described below. sub-divided in two areas that receive the infiltrated water from rainfall.1 𝑃𝑎𝑟𝑒𝑎 = 1 − (𝑃𝑐𝑡𝑖𝑚 + 𝐴𝑑𝑖𝑚𝑝) (J. The pervious capacity (or fraction) of this area will decrease as the rainfall grows.5) with T: relative content of Upper Zone Tension Water [-]. UztwC: the Upper Zone Tension Water storage [L]. The water can be only consumed by evapotranspiration.3) 𝑈𝑧𝑓𝑤𝐶∙𝑈𝑧𝑡𝑤𝑀𝑎𝑥−𝑈𝑧𝑡𝑤𝐶∙𝑈𝑧𝑓𝑤𝑀𝑎𝑥 𝑈𝑧𝑡𝑟 = 𝑑 ( 𝑈𝑧𝑓𝑤𝑀𝑎𝑥+𝑈𝑧𝑡𝑤𝑀𝑎𝑥 𝑈𝑧𝑡𝑟 = 0 ) /𝑑𝑡 if 𝑇 < 𝐹 (J. Upper Zone Tension Water storage This storage represents the water bound by adhesion and cohesion in between the soil pores as well as the water intercepted by vegetation.1) with Parea : permanent pervious area fraction of the basin [-] . Uztr: infiltration transfer from the Free Water storage to the Tension Water storage [L/T]. The water that reaches this reservoir comes from the rain felt into the pervious area of the basin. the SAC model divides the soil in 2 zones: - An Upper Zone. Adimp : maximum fraction of an additional impervious area due to saturation [-]. The expressions of these ratios and the infiltration transfer are shown in equations J.2 to J. depending on the storage of the permanent pervious and impervious areas. Pctim : permanent impervious area fraction [-] . sub-divided in 3 reservoirs.

7) (J. The first 2 factors define the percolation demand from the Lower Zone (DDA). The water may be depleted by evapotranspiration. 𝑃𝐵𝑎𝑠𝑒 = 𝐿𝑧𝑓𝑝𝑀𝑎𝑥 ∙ 𝐿𝑧𝑝𝑘 + 𝐿𝑧𝑓𝑠𝑀𝑎𝑥 ∙ 𝐿𝑧𝑠𝑘 𝐷𝐸𝐹𝑅 = (𝐿𝑧𝑡𝑤𝑀𝑎𝑥 − 𝐿𝑧𝑡𝑤𝐶) + (𝐿𝑧𝑓𝑝𝑀𝑎𝑥 − 𝐿𝑧𝑓𝑝𝐶) + (𝐿𝑧𝑓𝑠𝑀𝑎𝑥 − 𝐿𝑧𝑓𝑠𝐶) 𝐿𝑧𝑡𝑤𝑀𝑎𝑥 + 𝐿𝑧𝑓𝑝𝑀𝑎𝑥 + 𝐿𝑧𝑓𝑠𝑀𝑎𝑥 𝐷𝐷𝐴 = 𝑃𝐵𝑎𝑠𝑒 ∙ (1 + 𝑍𝑝𝑒𝑟𝑐 ∙ 𝐷𝐸𝐹𝑅𝑅𝑒𝑥𝑝 ) (J. the percolation to the Lower Zone is computed prior to the interflow computation as it is its preferred path. UzfwC: the Upper Zone Free Water storage [L]. UzfwMax: the Upper Zone Free Water capacity [L]. The water that reaches this reservoir comes from the infiltration of the Upper Zone Tension Water storage The following concepts are defined:  The Upper Zone Free Water Maximum capacity (UzfwMax) is the maximum Free Water storage admissible in the Upper Zone. percolation to the Lower Zone or horizontal flow (surface runoff and interflow). The effective percolation depends on the percolation demand and the ratio between the Upper Zone Free Water content and the Upper Zone Free Water capacity as shown in 1 http://www. It depends on the deficiency of the Lower Zone moisture volumes. Zperc: ratio of maximum and minimum percolation rates [-].pdf RS MINERVE – Technical Manual Page 39/94 . On the one hand.nws. DEFR: relative Lower Zone water deficiency [-].6) (J. Rexp: shape parameter of the percolation curve [-]. This demand represents the percolation in case of total percolation availability on the Upper Zone Free Water storage. when the Upper Zone Tension Water volume has been filled.noaa. DDA: Lower Zone maximum percolation demand [L/T]. this free water is rapidly transformed into Upper Zone Tension Water until tension water relative content requirements are satisfied. Although free water is present during filling of Upper Zone Tension Water (by infiltration).  The Interflow depletion rate (Uzk) represents the portion of the Upper Zone Free Water that is transferred outside as interflow.Chapter 2: Hydrological models description [L].gov/oh/hrl/nwsrfs/users_manual/part2/_pdf/23sacsma. excess moisture above the Upper Zone Tension Water capacity is accumulated in the Upper Zone Free Water.  The Upper Zone Free Water Content (UzfwC) is the Free Water stored in the Upper Zone at any given time.1 In addition. on the soil properties.8) with PBase: the continuing percolation rate under saturated condition (maximal percolation) [L/T]. and on the water relative storage in this reservoir. Upper Zone Free Water storage This storage represents the water not attached to the soil particles and free to move under gravitational forces.

UzfwC: Upper Zone Free Water content [L]. The following concepts are defined:   The Lower Zone Tension Water Maximum capacity (LztwMax) is the maximum Tension Water storage admissible in the Lower Zone. Moreover. One fraction (Pfree) of this percolation goes directly to the Free Water storages. the interflow occurs only when the rate of precipitation exceed the rate at which downward motion can occur from the Upper Zone Free Water. On the other hand.10 and J. It characterizes the volume of moisture in the lower soils which will be claimed by dry soil particles when moisture from a wetting front reaches that depth. even if the Tension Water storage is not yet full. UzfwMax: Upper Zone Free Water capacity [L].Chapter 2: Hydrological models description equation J.11) with PERC: real percolation [L/T]. as shown in equation J.11 respectively): 𝑃𝐸𝑅𝐶 = 𝐷𝐷𝐴 ∙ 𝑈𝑧𝑓𝑤𝐶 𝑈𝑧𝑓𝑤𝑀𝑎𝑥 (J. Parea : pervious area fraction of the basin [-]. RS MINERVE – Technical Manual Page 40/94 . proportionally to the remaining ETP of the Upper Zone Tension Water storage. the excess precipitation results in surface runoff QSurfRunoff. The water is consumed through evapotranspiration. A: surface of the basin [L2]. 2 If the Tension Water storage of the Lower Zone is completely full. Lower Zone Tension Water storage This storage represents the semi-saturated zone.9. The water reaching this reservoir comes from the Upper Zone percolation. UzfwC: Upper Zone Free Water content [L]. The Lower Zone Tension Water Content (LztwC) is the Tension Water stored in the Lower Zone at any given time. then percolation goes entirely to the 2 Free Water storages. DDA: Lower Zone maximum percolation demand [L/T]. and the other fraction (1-Pfree) goes to the Tension Water storage of the Lower Zone 2.10) 𝑃𝐸𝑅𝐶 ≤ 𝑑[(𝐿𝑧𝑡𝑤𝑀𝑎𝑥 − 𝐿𝑧𝑡𝑤𝐶) + (𝐿𝑧𝑓𝑝𝑀𝑎𝑥 − 𝐿𝑧𝑓𝑝𝐶) + (𝐿𝑧𝑓𝑠𝑀𝑎𝑥 − 𝐿𝑧𝑓𝑠𝐶)]/𝑑𝑡 (J. In addition.12) 3 with Qinterflow: interflow [L /T]. Uzk: interflow depletion rate from the Upper Zone Free water storage [T-1].12: 𝑄𝑖𝑛𝑡𝑒𝑟𝑓𝑙𝑜𝑤 = 𝑈𝑧𝑘 ∙ 𝑈𝑧𝑓𝑤𝐶 ∙ 𝑃𝑎𝑟𝑒𝑎 ∙ 𝐴 (J. the effective percolation is limited by the humidity content in the Upper Zone and also by the deficiency in the Lower Zone (equations J.The water also reaches this storage from transfers between the other Lower Zone Free Water storages. It depends on the water content and on its interflow depletion rate Uzk.9) 𝑃𝐸𝑅𝐶 ≤ 𝑑(𝑈𝑧𝑓𝑤𝐶)/𝑑𝑡 (J. when the Upper Zone Free Water storage is completely full and the precipitation intensity exceeds both the percolation rate and the maximum interflow drainage capacity.

The Lower Zone supplementary Free Water Maximum capacity (LzfsMax) is the maximum supplementary water storage admissible in the Lower Zone. F: relative content of Lower Zone Free Water [-].16) (J. There are two types of Lower Zone Free Water: a primary type with a very slow draining providing baseflow over long periods of time. The water that reaches these storages comes from the Pfree fraction of the percolation of the Upper Zone Free Water. in the whole Lower Zone [-]. the aquifer.17) with DEL: the Free-Tension Water transfer in the Lower Zone [L/T]. susceptible to become baseflow [L]. and a supplementary type which supplements the baseflow after a period of relatively recent rainfall. i.e. If the relative content of the Lower Zone Tension Water is smaller than the relative fullness of the global Lower Zone. Lower Zone Primary and Supplementary Free Water storages This storage represents the saturated zone of the subsoil. This fraction is distributed into the primary and supplementary storages according to their deficiency of water.13) 𝑇= 𝐿𝑧𝑡𝑤𝐶 𝐿𝑧𝑡𝑤𝑀𝑎𝑥 (J. susceptible to become baseflow [L]. Rserv: fraction of Lower Zone Free Water not transferable to Lower Zone Tension Water[-].14) 𝐹= 𝐿𝑧𝑓𝑝𝐶 + 𝐿𝑧𝑓𝑠𝐶 − 𝑅𝐹𝑊 𝐿𝑧𝑓𝑝𝑀𝑎𝑥 + 𝐿𝑧𝑓𝑠𝑀𝑎𝑥 − 𝑅𝐹𝑊 (J. the incoming excess is transferred from Free Water to Tension Water. The Lower Zone primary Free Water Content (LzfpC) is the primary Free Water stored in the Lower Zone at any given time. a water transfer DEL occurs from the Lower Zone Free Supplementary reservoir to the Lower Zone Tension Water reservoir. RFW: Lower Zone Free Water not transferable to Lower Zone Tension Water. The following concepts are defined:     The Lower Zone primary Free Water Maximum capacity (LzfpMax) is the maximum primary water storage admissible in the Lower Zone. The Depletion rate of the Lower Zone primary Free Water storage (Lzpk) is the portion of primary Free Water that drains as baseflow per day.13 to J.Chapter 2: Hydrological models description If the relative content of Tension Water (ratio between its content and its maximum storage) is smaller than the relative content of Free Water. R: ratio between the available water for evapotranspiration and the total water content.15) with T: relative content of Lower Zone Tension Water [-]. RS MINERVE – Technical Manual Page 41/94 .15: 𝑅𝐹𝑊 = 𝑅𝑠𝑒𝑟𝑣 ∙ (𝐿𝑧𝑓𝑝𝑀𝑎𝑥 + 𝐿𝑧𝑓𝑠𝑀𝑎𝑥) (J. RFW: Lower Zone Free Water not transferable to Lower Zone Tension Water. as described by the following equations: 𝑅= 𝐿𝑧𝑓𝑝𝐶 + 𝐿𝑧𝑓𝑠𝐶 + 𝐿𝑧𝑡𝑤𝐶 − 𝑅𝐹𝑊 𝐿𝑧𝑓𝑝𝑀𝑎𝑥 + 𝐿𝑧𝑓𝑠𝑀𝑎𝑥 + 𝐿𝑧𝑡𝑤𝑀𝑎𝑥 − 𝑅𝐹𝑊 𝐷𝐸𝐿 = 𝐿𝑧𝑡𝑤𝑀𝑎𝑥 ∙ 𝑑(𝑅 − 𝑇)/𝑑𝑡 (J. The expressions of both ratios are shown in equations J.

24): 𝐸𝑇𝑅1 = 𝐸𝑇𝑃 ∙ 𝑈𝑧𝑡𝑤𝐶 ≤ 𝑈𝑧𝑡𝑤𝐶 𝑈𝑧𝑡𝑤𝑀𝑎𝑥 (J. that lead to filtrations throughout the aquifer. QSubSurf: discharge lost into the aquifer [L3/T].21: 𝑄𝑆𝑢𝑏𝑆𝑢𝑟𝑓 = 𝑆𝑖𝑑𝑒 ∙ 𝑄𝑏𝑎𝑠𝑒𝑓𝑙𝑜𝑤 (J. The drained water leaving these two storages (primary flow Qprimary and supplementary flow Qsupplementary) follows the Darcy's law and forms the baseflow of the Lower Zone QLower Zone as shown in equations J.18 to J. Parea : pervious area fraction of the basin [-].21) with Side: ratio of deep percolation from Lower Zone Free Water storages [-].22) (J. etc. the evapotranspiration is proportional to its content (equation J. called subsurface flow QSubSurf.Chapter 2: Hydrological models description  The Lower Zone supplementary Free Water Content (LzfsC) is the supplementary Free Water stored in the Lower Zone at any given time. from riverside vegetation and from impervious areas. And so this Side parameter is used to correct the baseflow as follows: 𝑄𝑏𝑎𝑠𝑒𝑓𝑙𝑜𝑤 = 𝑄𝐿𝑜𝑤𝑒𝑟 𝑍𝑜𝑛𝑒 1 + 𝑆𝑖𝑑𝑒 𝑄𝑆𝑢𝑏𝑆𝑢𝑟𝑓 = 𝑆𝑖𝑑𝑒 ∙ 𝑄𝐿𝑜𝑤𝑒𝑟 𝑍𝑜𝑛𝑒 1 + 𝑆𝑖𝑑𝑒 (J. LzfpC: Lower Zone primary Free Water Content [L]. QLower Zone: total baseflow produced in the Lower Zone [L3/T]. Qsupplementary: supplementary baseflow [L3/T]. Lzsk: depletion rate of the Lower Zone supplementary Free Water storage [T-1]. fractured rocks.20: 𝑄𝑝𝑟𝑖𝑚𝑎𝑟𝑦 = 𝐿𝑧𝑝𝑘 ∙ 𝐿𝑧𝑓𝑝𝐶 ∙ 𝑃𝑎𝑟𝑒𝑎 ∙ 𝐴 (J. LzfsC: Lower Zone supplementary Free Water Content [L].24) with ETP: potential evapotranspiration [L/T]. The remaining evapotranspiration demand RED and the updated content of water in the Upper Zone Tension Water reservoir UztwC are then calculated: RS MINERVE – Technical Manual Page 42/94 . Qbaseflow: Lower Zone Free Water volume actually integrated to the channel outflow hydrographs [L3/T].20) with Qprimary: primary baseflow [L3/T]. Lzpk: depletion rate of the Lower Zone primary Free Water storage [T-1]. A: surface of the basin [L2]. They are described hereafter.  The Depletion rate of the Lower Zone supplementary Free Water storage (Lzsk) is the portion of supplementary Free Water that drains as baseflow per day.  ETR1: evapotranspiration from the Upper Zone Tension Water reservoir. from the transfers between Free Water and Tension Water storages.23) Evapotranspiration The real evapotranspirations are obtained from each Tension Water storage. The model allows including baseflow losses.18) 𝑄𝑠𝑢𝑝𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑟𝑦 = 𝐿𝑧𝑠𝑘 ∙ 𝐿𝑧𝑓𝑠𝐶 ∙ 𝑃𝑎𝑟𝑒𝑎 ∙ 𝐴 (J. If this reservoir is not full. due to the existence of geological faults.19) 𝑄𝐿𝑜𝑤𝑒𝑟 𝑍𝑜𝑛𝑒 = 𝑄𝑝𝑟𝑖𝑚𝑎𝑟𝑦 + 𝑄𝑠𝑢𝑝𝑝𝑙𝑒𝑚𝑒𝑛𝑡𝑎𝑟𝑦 (J. The Side parameter captures these losses as shown in equation J.

34) with ETRtotal: total evapotranspiration [L3/T].31) ETR5: evapotranspiration from the impervious fraction of the basin.27) Then. but it is limited by the Upper Zone Free Water content UzfwC: 𝐸𝑇𝑅2 = min{𝑅𝐸𝐷. Parea: pervious area fraction of the basin [-]. The evapotranspiration availability corresponds to the remaining evapotranspiration demand in the Lower Zone. 𝑑(𝑈𝑧𝑓𝑤𝐶)/𝑑𝑡} (J.28) The variation of the water content UzfwC in the Upper Zone Free Water reservoir is defined as follows: 𝑑(𝑈𝑧𝑓𝑤𝐶)/𝑑𝑡 = −𝐸𝑇𝑅2  ETR3: evapotranspiration from the Lower Zone Tension Water reservoir. and inversely proportional to the Tension Water reservoirs' capacity of the Upper Zone and the Lower Zone: 𝐸𝑇𝑅3 = 𝑅𝐸𝐷𝐿𝑧 ∙  (J. This evapotranspiration is equal to the remaining RED left by the Tension Water reservoir.Chapter 2: Hydrological models description  𝑅𝐸𝐷 = 𝐸𝑇𝑃 − 𝐸𝑇𝑅1 (J.29) 𝐿𝑧𝑡𝑤𝐶 ≤ 𝑑(𝐿𝑧𝑡𝑤𝐶)/𝑑𝑡 𝑈𝑧𝑡𝑤𝑀𝑎𝑥 + 𝐿𝑧𝑡𝑤𝑀𝑎𝑥 ETR4: evapotranspiration from the basin covered by riverside vegetation. Each evapotranspiration value is weighted based on the portion of the basin area in which it is produced. The total evapotranspiration is finally: 𝐸𝑇𝑅𝑡𝑜𝑡𝑎𝑙 = [(𝐸𝑇𝑅1 + 𝐸𝑇𝑅2 + 𝐸𝑇𝑅3) ∙ 𝑃𝑎𝑟𝑒𝑎 + 𝐸𝑇𝑅4 ∙ 𝑅𝑖𝑣𝑎 + 𝐸𝑇𝑅5 ∙ 𝐴𝑑𝑖𝑚𝑝] ∙ 𝐴 (J. RS MINERVE – Technical Manual Page 43/94 . Riverside vegetation absorbs the deficiency of potential evapotranspiration from the Upper Zone and the Lower Zone: 𝐸𝑇𝑅4 = 𝐸𝑇𝑃 − 𝐸𝑇𝑅1 − 𝐸𝑇𝑅2 − 𝐸𝑇𝑅3  (J. VETR1: evapotranspirated water from the Upper Zone Tension Water reservoir in a time step [L].32) 𝑡2 𝑉𝐸𝑇𝑅1 = ∫ 𝐸𝑇𝑅1. the remaining evapotranspiration REDLz from the Lower Zone is: 𝑅𝐸𝐷𝐿𝑧 = 𝑅𝐸𝐷 − 𝐸𝑇𝑅2 (J. Evapotranspiration in this reservoir is proportional to the remaining evapotranspiration and to the Lower Zone Tension Water content.33) 𝑡1 with AdimC: sum of the Upper Zone and Lower Zone Tension Water Content in the Adimp area [-]. Pctim: permanent impervious area fraction of the basin [-]. Riva: riparian vegetarian area fraction [-].26) ETR2: evapotranspiration from the Upper Zone Free Water reservoir. 𝑑𝑡 (J. 𝐸𝑇𝑅5 = 𝐸𝑇𝑅1 + (𝑅𝐸𝐷 + 𝐸𝑇𝑅2) ∙ 𝐴𝑑𝑖𝑚𝐶 − 𝑈𝑧𝑡𝑤𝐶 − 𝑉𝐸𝑇𝑅1 𝑈𝑧𝑡𝑤𝑀𝑎𝑥 + 𝐿𝑧𝑡𝑤𝑀𝑎𝑥 (J.30) (J.25) 𝑑(𝑈𝑧𝑡𝑤𝐶)/𝑑𝑡 = −𝐸𝑇𝑅1 (J.

nws. A more detailed guide about the SAC-SMA equations can be found on the following National Weather Service link: http://www. A: surface of the basin [L2].noaa.gov/iao/iao_SAC_SMA.php RS MINERVE – Technical Manual Page 44/94 .Chapter 2: Hydrological models description Adimp: maximum fraction of an additional impervious area due to saturation [-].

3) 0 The term b represents the cross sectional variation for the level  and constitutes an integration variable according to Figure 11. The three possibilities are presented hereafter and their parameters in Table 11.2) with A: cross sectional flow area [L2]. Venant equations solving the 1D unsteady flow are: A Q  0 t x (K. J0: bottom slope. Muskingum-Cunge or Kinematic wave.1)  Q   Q 2    g ·I1   g ·A·J 0  J f   g ·I 2 t x  A  (K. the equation defining I1 is reduced to: RS MINERVE – Technical Manual Page 45/94 .Chapter 2: Hydrological models description 2. Venant only) Initial discharge Regular Range >0 >0 0.12 Channel routing description The channel routing can be solved by the St-Venant. I2: coefficient for cross sectional variation [L2]. The term I1 takes into account the shape of the transversal profile and is calculated as follows: h I1   (h   )  b( ) d (K.2 ensures the conservation of momentum. I1: profile coefficient [L3]. St. Venant equations can be solved in RS MINERVE for a trapezoidal profile (Figure 12): 1 Figure 11 Descriptive sketch for parameters Figure 12 Transversal profile available for the used in the calculation of I1 computation of channel routing For this trapezoidal section. Table 11 List of parameters for the channel routing Object Name L B0 m J0 K nSec QIni Reach Units m m m1/3/s m3/s Description Length Width of the channel base Side bank relation coefficient (1H/mV) Slope Strickler coefficient Number of sections (St. VENANT routing The St. Jf: friction slope [-]. Equation K.1 to 1 >0 10 to 90 >0 - ST.1 expresses the mass conservation while equation K. Q: discharge [L3/T].

∆x: longitudinal increment [L]. j 1  g  I1.9) n 2  J 0  t  g  A j 1  J f n n j 1 with index j and j+1 representing the spatial position. Applicable solutions are the diffusive and kinematic waves which are presented in the following. I2 is: h I 2   (h   ) 0 b x h (K.1 and K. g: gravity constant [L2/T].7) For a prismatic channel. In practice.Chapter 2: Hydrological models description B  h 2 m  h3  2 3 I1  (K. equations K. according to the physical situation to be simulated.6) with Jf: friction slope [-]. m: side bank relation coefficient (1 vertical / m horizontal) [-]. A: flow area [L2]. The term I2 takes into account the variation of the section along the channel. some terms of the complete dynamic equations can be eliminated in order to get simplified expressions without losing precision. In the case of a prismatic channel I2 is equal to zero.2 are solved by the Euler method (first order) as follows: A j 1 n 1  A j 1  n 1  Q j 1 n t n n (Q j 1  Q j ) x n Q j 1 n n 2 2 t (Q j 1 ) (Q j ) n n  (   g  I1. In general. The downstream boundary condition used by RS MINERVE is the normal flow depth.8) (K. The friction slope Jf is calculated according to Manning-Strickler: Jf  Q | Q | 4 A 2  K 2  Rh 3 (K. Rh : hydraulic radius [L]. RS MINERVE – Technical Manual Page 46/94 . Lp : wetted perimeter [L]. j )  n n x A j 1 Aj ( A j  A j 1 ) n  t  g  (K. exponent n and n+1 representing the time increment.5) Rh  A Lp (K. K: Strickler coefficient [L1/3/T].4) with B: width of the base of the transversal profile [L]. h: water level [L].

15) This equation is called « equation of the kinematic wave » and describes the simple convection of the flow with a velocity c according to equation K.15 yields:  X Qj n 1    Q j  1  X  Q j 1 ct n RS MINERVE – Technical Manual n 1  Q j 1 n   2 Q 1 n 1 j 1  Qj n 1   12 Q x n j 1  Qj n  0 (K.11 is reduced to: Q  dQ  Q  0  t  dA  x0 x (K. 1991) it is possible to express equation K.2 as follows: Q  Q dD  Q D 2  2Q   0  t  BD dh  x 2 B Q x 2 (K.14): Q dD BD dh (K.11 is an equation with partial derivatives of parabolic type which represents the convection and the diffusion of the variable Q.Chapter 2: Hydrological models description Routing according to Muskingum-Cunge Ruling out the first two terms of equation K. D: discharge rate [L3/T].13.16) t    1 1 n 1 n 1 n n Q  Q j  Q j 1  Q j Q 2 j 1 2  x x  (K.12) 2 Equation K.14) c Based on the hypothesis of a clearly defined relation between the flow Q and the water level h.10) This new equation corresponds to the approximation of the diffusive wave.17) Applying this scheme to equationK.13) and diffused with a diffusion coefficient  (equation K. equation K.18) Page 47/94 .11) with B: width of the bottom of the transversal profile [L].2 yields: I1  A  J 0  J f   I 2 x (K. the flow transported with a velocity c (equation K. With the supplementary hypothesis of a prismatic channel (Cunge. The discharge rate is the capacity of a cross section of a channel to transport a certain flow and is defined as: Q  DJ0 1 (K.13) D2  2B  Q (K. Hence. It can be solved by the following numerical finite difference scheme:  Q X Q j  t n 1  Qj n   1  X Q n 1 j 1  Q j 1 n  (K.

1980) corresponding to the name of the river localized in the United States where the method was employed for the first time.24) X corresponds to: X  1 D3  2 2x Q Q dD dh (K.24) According to this analysis proposed by Cunge (1969) it can be recognized that the Muskingum equation is a solution in terms of finite differences of the equation of the diffusive wave (K. n) assuming x/t = c and neglecting the quadratic terms (x2) the equation can be written as follows: Q t c Q x  cB 2Q x 2  0 1  B  x   X  2  (K.Chapter 2: Hydrological models description We can express the solution of this equation as a function of the unknown variable Q j+1n+1 namely: Q j 1 n 1  C1Q j n 1  C 2 Q j  C3Q j 1 n n (K.20) t 2 t K 1  X   2 C3  t K 1  X   2 K 1  X   x c K  (K. This is not only an appearance since developing the terms of equation K.23) (K. K is defined by equation K.25) This function of the diffusive wave implemented at present in RS MINERVE is capable of solving the Muskingum-Cunge equation for the trapezoidal geometry of a transversal profile according to Figure 12.21) Q j 1  Q j n A j 1  A j n n c n (K.19) with: C1   C2  KX  t 2 K 1  X   KX  t 2 t 2 (K. hydraulic engineers might recognize the equation of Muskingum (Boillat. RS MINERVE – Technical Manual Page 48/94 .22) Here.18 in terms of a Taylor series around the point (j.21 and according to the celerity c (equation K. The Muskingum method represents an approximation by finite differences of the equation for the kinematic wave.11) under the condition of correctly introducing the value of the parameters K and X.

since high discharges are transferred downstream more rapidly than small ones. The equation of the cinematic wave as presented in the previous chapter is presented as follows: Q  dQ  Q  0  t  dA  x0 x (K. Since no diffusive term appears in the equation the peak discharge remains constant and is not reduced. the general behaviour of a flood is modified. The initial parameters are identical to those of the model of the diffusive wave.29) n 1    Q j  (1   )  Q j 1 n 1  n 1   Q j  (1  n 1  RS MINERVE – Technical Manual )Qj n si   1 n 1 si   1 (K. no reduction of the flood is produced as mentioned previously.28) t L (K. Figure 12). The solution is performed according to the following equations: Q j 1  Q j n A j 1  A j n n c n   c Q j 1 Q j 1 (K.Chapter 2: Hydrological models description Routing according to the kinematic wave The kinematic wave model is the simpler routing model where the terms of inertia and pressure of the St. Venant equations are supposed to be negligible. As a consequence. Opposite to the model of the diffusive wave. the cinematic hypothesis supposes that the gravity forces are identical. This implies that there is an explicit relationship between the flow and the water level (measured normal water depth).30) Page 49/94 .26) This is a simple equation of convection which indicates that the flow Q is transported downstream with a celerity c which is defined as: c Q A (K. On the contrary. The geometry for the transversal profiles also corresponds to the same than for the MuskingumCunge method (trapezoidal channels. though with an opposite sign.27) This rather simple model transports each point of the hydrograph from upstream to downstream with a velocity c. to the friction forces.

it is necessary to know the outflow as a function of the water volume  in the reservoir.13 Reservoir The transient evolution of a water volume  in a reservoir is described by the following retention equation: d  Qe  Qs dt (L.m3 masl Description Level .Chapter 2: Hydrological models description 2. To solve equation L.Volume relation Initial level in the reservoir Page 50/94 . Table 12 Reservoir paired data and initial condition required Object Reservoir Name H-V (paired data) Hini RS MINERVE – Technical Manual Units masl . Qe: inflow in the reservoir [L3/T]. Usually. This operation is possible if relation between the water level and the water volume is known (Table 12).1) with  : volume in the reservoir [L3]. Qs: outflow [L3/T].1. reservoirs are equipped with turbines. pumps and spillways depending on the water level in the reservoir.

i  ( H i 1  H n ) n Qoutflow. Hence.Discharge relation Page 51/94 . Hi: reservoir water level [L].Chapter 2: Hydrological models description 2.m3/s Description Level . Table 13 HQ paired data required Object HQ Name H-Q (paired data) RS MINERVE – Technical Manual Units masl .1) with Qoutflown: discharge at instant n [L3/T]. Qoutflow. the discharge in then calculated as follows: Qoutflow  0 if H n  H1 n Qoutflow  Qdev.i H i 1  H i if H i  H n  H i 1 (M.14 Level-Discharge Relation (HQ) The HQ object provides an outflow depending on a level in a reservoir. Hn: water level at instant n [L].i: discharge flow for a water level Hi [L3/T]. The outflow is calculated by the help of a « Level – Discharge » relation (Table 13).i 1  Qoutflow.

3.% (paired data) Zcentral masl L m Hydropower D m K m ν m2/s Default Price euro/Kwh Description Discharge-Performance relation Hydropower plant altitude Length of the pipe Diameter of the pipe Roughness Kinematic viscosity Default price. In addition. Re. 2.1 to N.2) Revenuen  Power n  Price n (N.5. Pricen: Price of the energy at instant n [€/Kwh].1) (N. the Reynolds number. only used if no data exists in the database A default price can be introduced for first approximation of the revenue. but also the total energy produced and the total obtained revenue. As outputs. 7  Re    1 RS MINERVE – Technical Manual (N. 2 Z net  ( Z water n n 8  L  Qn  Z central )  f  g   2  D5 Power n   n 1000  Qn  g  Z net n (N. the equation of Colebrook-White (Colebrook and White.3) with Hnetn: net height at instant n [L]. For this calculation. Zwatern: water height in the reservoir at instant n [L].51  k/D  2  log10    3 .Chapter 2: Hydrological models description 2.15 Hydropower The inputs of this object are a reservoir level and a discharge. Colebrook. If the discharge actually provided to the object Hydropower is higher than the range proposed in the relation Q-η of this object. the altitude (Zcentral) of the hydropower plant. is presented in equation N. 9. Qn: discharge at instant n [L3/T]. where λ represents f. It also needs the paired data relation « Discharge (Q) – Performance (η) » for the turbine. D: diameter [L]. the maximum discharge of the paired data Q-η is taken for the energy production calculations. as presented in equations N. Zcentral: hydropower plant altitude [L].4) Page 52/94 . Simmons. it takes into account the head loss in terms of volumetric flow rate in a full-flowing circular pipe from the Darcy-Weisbach equation (Darcy. It is presented in equation N. f: friction factor [-]. η: performance of the turbine at instant n [%]. 1937. the length (L). the object calculates not only the power and the revenue for each time step. 2008). the diameter (D) and the roughness (K) of the pipe as well as the kinematic viscosity of the fluid (υ). It is used only if the object does not have a price series in the database. Powern: power at instant n [Watt]. 1939) is used.81 [L2/T]. Revenuen: Revenue of the turbine at instant n [€/Kwh]. For the calculation of the friction factor f of the Darcy-Weisbach equation. g: gravity. 1857. Table 14 Diversion paired data required Object Name Units Q-η m3/s .4. L: length of the pipe [L].

Chapter 2: Hydrological models description Re  v  D  (N. ν: kinematic viscosity [L2/T] The equation N.4 is solved for the range of discharges of the paired data Q-η.5) with λ: friction factor [-]. v: velocity [L/T]. Re: Reynolds number [-]. k: roughness [L]. RS MINERVE – Technical Manual Page 53/94 . providing the paired data relation « Discharge (Q) – Friction factor (λ) » for user information.

With this information.Diverted flow relation Page 54/94 . the Diversion calculates the diverted hydrograph and the downstream hydrograph.1: Qup  Qdown  Qdiverted n n n (O.16 Diversion This object needs the paired data relation « Inflow – Diverted flow » as information as well as the incoming hydrograph.1) with Qupn: total flow upstream at instant n [L3/T].m3/s Description Upstream flow .Chapter 2: Hydrological models description 2. Qdivertedn: diverted flow at instant n [L3/T]. The relation describes the behaviour of the diversion and is generated by the user. as presented in equation M. Table 15 Diversion paired data required Object Diversion Name Qup-Qdiverted (paired data) RS MINERVE – Technical Manual Units m3/s . Qdownn: downstream flow at instant n [L3/T]. who performs a calculation for the behaviour of the diversion and then creates the relation « Inflow – Diverted flow ».

Chapter 2: Hydrological models description 2. At the same time. With this information. Qdownn: downstream flow at instant n [L3/T].1) Qconsumed  Qup n Qdown  0 n n n Qshorfall  Qdemand  Qup n n if Qdemand  Qup n n with Qconsumedn: consumed discharge at instant n [L3/T]. the object also calculates the shortfall discharge series during the simulation period. only used if no data exists in the database Page 55/94 . as presented in equation P. Qdemandn: demanded consummation at instant n [L3/T].17 Consumer A series in the database is used as the consummation demand of the Consumer object. the parameter “Default QDemand” is used for the whole period of the simulation as uniform demand. If no information about consummation exists in the database. Qconsumed  Qdemand n n n Qdown  Qup  Qdemand n Qshorfall  0 n n if Qdemand  Qup n n (P. the Consumer calculates the consumed discharge as well as the downstream hydrograph. Qupn: total flow upstream at instant n [L3/T].1. Table 16 Consumer optional parameter Object Consumer Name Units Default QDemand m3/s RS MINERVE – Technical Manual Description Default demand of consummation.

1) Qlost  Qup  (1  Efficiency ) (Q.18 Structure efficiency This object needs an efficiency coefficient as information to describe the efficiency of a structure such a canal or a pipe. Qdownn: downstream flow at instant n [L3/T].1 and Q. as well as the incoming hydrograph. An efficiency of 0 generates a complete loss of the input.Chapter 2: Hydrological models description 2. Qlostn: lost flow at instant n [L3/T]. Table 17 Structure efficiency parameter required Object Structure Efficiency Name Units Efficiency - RS MINERVE – Technical Manual Description Efficiency of the structure Page 56/94 . An efficiency of 1 provides an output flow equivalent to the input hydrograph. The downstream hydrograph is calculate as presented in equation Q.2) n n n n with Qupn: total flow upstream at instant n [L3/T].2 : Qdown  Qup  Efficiency (Q.

in tabular form « t [s] – value [depending on the series] ». temperature or ETP can be directly introduced.20 Time series In this object.(depending on the series) Description Time – Value series Page 57/94 . precipitation. temporary series of flow.Chapter 2: Hydrological models description 2. The time is incorporated in seconds and the associated values in their corresponding units. Table 18 Time series required data Object Time Series Name Series (paired data) RS MINERVE – Technical Manual Units s .

21 Sensor This object can be linked to any other object and send information to the states of a regulation each time a threshold is exceeded or values decrease below a threshold. Table 19 Sensor required data Object Name Units Description Sensor Thresholds (depending on the linked object) Thresholds RS MINERVE – Technical Manual Page 58/94 .Chapter 2: Hydrological models description 2. The threshold values are specified by user in the object.

Table 20 Regulation and States required data Object Regulation/States Name Transitions Units - Description Operators and thresholds The possible operators to use (alone or combined) are presented in Table 21. depending on operators and defined thresholds from sensors (Table 20). inside every State. Inside this object. etc). several « States » can be created. transitions between States are also introduced.22 Regulation and State When a regulation is selected. Once all objects are correctly linked. pump. Table 21 Operators for states transition rules Operator >= than < than and >= than and < than or >= than or < than RS MINERVE – Technical Manual Definition Bigger than or equal to Smaller than …and "bigger than or equal to" …and "smaller than" …or "bigger than or equal to" …or "smaller than" Page 59/94 . diversion. A typical example of regulation is the implementation of a turbine / pump law as a function of the water level in the reservoir. a sub-model is automatically generated. The choice about which state is active is carried out as a function of the thresholds information (from the Sensors). several objects are introduced as a function of the demands (turbine.Chapter 2: Hydrological models description 2. Thus.

It is defined as presented in Eq.t )) 2 ref . tf Nash  ln  1   (ln( Q t ti sim .1 Nash coefficient The Nash-Sutcliffe criteria (Nash and Sutcliffe.t )  ln( Q ref )) 2 with Nash-ln: Nash-Sutcliffe coefficient for log values [-].. Schaefli and al.t IND. It varies from -∞ to 1. 2011). RS MINERVE – Technical Manual Page 60/94 .Chapter 3: Performance indicators Chapter 3: Performance indicators The Comparator object provides seven indicators values.2 tf  (ln( Q t ti )  ln( Qref .1..t ) 2  (Q  Q ref ) sim. It is defined as presented in Eq. with 1 representing the best performance of the model. 2005. Jordan. Qref.2. 3.t: simulated discharge at time t [L3/T].t t t i IND. presented hereafter... 2011). 2007. Qsim. IND. 2005. 3. 2004. It varies from -∞ to 1.1 2 with Nash: Nash-Sutcliffe model efficiency coefficient [-]. Nóbrega et al. 2009.t t t i tf ref . with 1 representing the best performance of the model and zero the same performance than assuming the average of all the observations at each time step. tf Nash  1   (Q  Qref . García Hernández et al. 1970) is used to assess the predictive power of hydrological models (Ajami et al. Viviroli et al. Q ref: average observed discharge for the considered period [L3/T].. IND.2 Nash coefficient for logarithm values The Nash-Sutcliffe coefficient for logarithm flow values (Nash-ln) is used to assess the hydrological models performance for low flows (Krause et al.t : observed discharge at time t [L3/T].

4 Bias Score The Bias Score (BS) is a symmetric estimation of the match between the average simulation and average observation (Wang et al.t  Q ref ) 2 IND.t ) 2 n IND. with 1 representing the best performance. It varies from 0 to 1..5 Relative Root Mean Square Error The Relative Root Mean Square Error (RRMSE) is defined as the RMSE normalized to the mean of the observed values (Feyen et al. the better the model performance is. 2005. with 1 representing the best performance of the model.4 with BS: Bias Score [-]. It varies from 0 to +∞.5. Q sim: average simulated discharge for the considered period [L3/T].4. t t ti sim.3 Pearson Correlation Coefficient The Pearson correlation coefficient shows the covariability of the simulated and observed discharges without penalizing for bias (AghaKouchak and Habib. 3. n: number of values [-]. Wang et al. IND. 2010.3 t ti with Pearson: Pearson Correlation Coefficient [-]. 2011).t  Qref .Chapter 3: Performance indicators 3.. tf Pearson   (Q tf tf  (Q t ti  Q sim )  (Qref . It varies from -1 to 1. It is defined as presented in Eq. 2011).   Q Q BS  1  max( sim . 3. Heppner et al. The smaller RRMSE..5 Q ref with RRMSE: relative RMSE [-].. IND. ref )  1 Q ref Q sim   2 IND. RS MINERVE – Technical Manual Page 61/94 ..3. tf  (Q t ti RRMSE  sim. El-Nasr et al. 2000. 2006) and is presented in Eq.t  Q ref ) sim. IND. t  Q sim ) 2   (Qref . It is defined as presented in Eq.

2007).Chapter 3: Performance indicators 3.8 tf Rmax   Qref .7 Normalized Peak Error The Normalized Peak Error (NPE) indicates the relative error between the simulated and the observed flow peaks (Masmoudi and Habaieb. 3.t ) with RVB: relative volume bias between forecast and observation for the considered period [-]. NPE  Smax  Rmax Rmax IND. The NPE varies from -∞ to +∞. An index near to zero indicates a good performance of the simulation. sometimes called differently.t ) sim. It is computed according to IND.9. Negative values are returned when simulated peak discharge is below the observed one. RS MINERVE – Technical Manual Page 62/94 . 2007.9 t ti with NPE: relative error between simulated and observed peak discharge [-]. 2000.t IND. while positive values mean the opposite. Smax : maximum simulated discharge for the studied period [L3/T]: Rmax : maximum observed discharge for the studied period [L3/T]. Gabellani and al.6 Relative Volume Bias The Relative Volume Bias (RVB. corresponds in this case to the relative error between the simulated and the observed volumes during the studied period (Ajami and al.t t ti IND.7 TO IND. in average. 2010) according to Eq. 1993. Sun and al. AghaKouchak and Habib. Schaefli and al. Negative values are returned when simulated discharge is. The RVB varies from -∞ to +∞. while positive values mean the opposite (overage model).. Moriasi et al. 2005.6. tf  (Q RVB   Qref . Values near to zero indicate a good performance of simulated peaks regarding observed ones.7 tf S max   Qsim.t t ti IND. 2004.6 tf  (Q t ti ref . IND. smaller than the average of the observed discharge (deficit model). 2004. Ajami and al.

Blasone et al. Variations of the Monte Carlo method are usually used in hydrological problem for parameterization optimization (Vrugt and al. 2006. so-called complex shuffling. 1993) based on a synthesis of the best features from several existing algorithms.. 1979) with the Rosenbrock algorithm (Rosenbrock. The first algorithm. 2006.. 1996. Rosenbrock algorithm have been also used for hydrological parameters optimisation (Abbot et Refsgaard. 2004.2 Objective function A flexible objective function (OF) has been developed for the module of calibration aiming to be adapted to the user’s requirements. Kamali et al... 2010).Chapter 4: Calibration of hydrological models Chapter 4: Calibration of hydrological models 4. RS MINERVE – Technical Manual Page 63/94 .. The SCE-UA method was designed for solving problems encountered in conceptual watershed model calibration (Hapuarachchi H.. et al. It has been modified to an Uniforme Adaptative Monte Carlo (UAMC) in this program to adjust the solution space after a defined group of simulations up to the convergence of the optimization. 1996) or optimization of numerical functions (Kang et al. 2008. combined the with other methods (van Griensven et al. Liu. 2001. 1992. 2004.. 2013). The third and last algorithm used in RS MINERVE is the Coupled Latin Hypercube and Rosenbrock (CLHR) It couples the Latin Hypercube algorithm (McKay et al. used since it can be interesting for solving complex problems in high dimensional spaces (Gilks et al. 2012).A. is a global optimization method (Duan et al. 2001). Jeremiah and al. The latin hypercube algorithm has been usually used in hydrology for sampling the initial parameter space. 1960).. This module uses an objective function defined by the user and different algorithms to solve it. 4. 2007). but has also been satisfyingly used in water resources management (Zhu et al. The indicators presented in Chapter 4 are used in this OF. 2011).. Lin et al. The second algorithm is a variation of the Adaptive Markov Chain Monte Carlo. generating a powerful tool for optimization of complex problems.. Ajami et al. and introduces the concept of complex information exchange.. the Shuffled Complex Evolution – University of Arizona (SCE-UA). including the genetic algorithm... each one weighted with a value defined by the user (Table 22). 2003. Wang et al.1 Introduction to the calibration module The module Calibrator of the RS Expert frame has been implemented for calibrating the parameters of the hydrological model. Muttil and Liong.P.

2011). NPE) since their ideal value corresponds to zero. and uses an adaptation of the Simplex Downhill search scheme (Nelder and Mead. to minimize the value or the absolute value for the last three indicators (RRMSE. 1965) to continuously evolve the population toward better solutions in the search space. The goal is to find a single best parameter set in the feasible space. at the same time. Step 2 Rank points: Sort the NPT points to increase criterion value so that the first point represents the point with the lowest criterion value and the last the one with the highest criterion value (assuming that the goal is to minimize the criterion value). use a uniform probability distribution to generate a sample. RVB. In the absence of prior information on the location of the global optimum.3 Shuffled Complex Evoluation – University of Arizona Model architecture The Shuffled Complex Evolution – University of Arizona (SCE-UA) method was developed to obtain the traditional best parameter set and its underlying posterior distribution within a single optimization run. Thus.1 4.Chapter 4: Calibration of hydrological models Table 22 Weights of the indicators for the objective function Indicator Weight Range of Values Ideal value Nash w1 -∞ to 1 1 Nash-ln w2 -∞ to 1 1 Pearson Correlation Coefficient w3 -1 to 1 1 Bias Score (BS) w4 0 to 1 1 Relative Root Mean Square Error w5 (RRMSE) 0 to +∞ 0 Relative Volume Bias (RVB) w6 -∞ to +∞ 0 Normalized Peak Error (NPE) w7 -∞ to +∞ 0 The OF is presented in Eq. OF  max( Nash  w1  Nashln  w2  Pearson  w3  BS  w4   RRMSE  w5  RVB  w6  NPE  w7 ) OF.1 and takes into account the ideal values of each indicator. progressively relinquishing occupation of regions with lower posterior probability (Mariani et al. 1994) and illustrated in Figure 13. the OF searches to maximize first four indicators (Nash. RS MINERVE – Technical Manual Page 64/94 .. Step 1 Generate sample: Sample NPT points in the feasible parameter space and compute the criterion value at each point. A general description of the steps of the SCE-UA method is given below (Duan et al. OF. Pearson and BS) since their ideal value is equal to the maximum possible value and. It starts with a random sample of points distributed throughout the feasible parameter space. Nash-ln..

Step 5 Shuffle complexes: Combine the points in the evolved complexes into a single sample population. return to Step 4. Step 4 Evolve each complex: Evolve each complex independently by taking NSPL evolution steps.Chapter 4: Calibration of hydrological models Step 3 Partition into complexes: Partition the NPT points into NGS complexes.NPG. Figure 15 illustrates how each evolution step is taken. RS MINERVE – Technical Manual Page 65/94 . and return to Step 4.. The complexes are partitioned in such a way that the first complex contains every NGS∙(k-1)+1 ranked point. the second complex contains every NGS*(k-1)+2 ranked point.. stop. otherwise. Step 6 Check convergence: If any of the pre-specified convergence criteria are satisfied. according to the Competitive Complex Evolution (CCE) algorithm. sort the sample population in order of increasing criterion value. Step 7 Check complex number reduction: If MINGS (the minimum number of complexes) < NGS... re-partition or shuffle the sample population into NGS complexes according to the procedure specified in the third step. and so on. each containing NPG points. where k = 1. continue. remove the complex with the lowest ranked points. If MINGS=NGS.2. set NGS=NGS-1 and NPT=NGS*NPG.

The contour lines in Figure 14 and Figure 15 represent a function surface having a global optimum located at (4.2). NGS=p. Figure 14b shows the locations of the points in the two independently evolved complexes at the end of the first cycle of evolution. It can be seen that one complex (marked by *) is converging towards the local optimum. Figure 14d illustrates the two complexes at the end of the second cycle of evolution. Each complex contains NPG (=5) points which are marked by • and * respectively. The two evolved complexes are shuffled according to step 5. It is clear that both complexes are now converging to the global optimum at the end of second cycle. Figure 14c displays the new membership of the two evolved complexes after shuffling. Figure 14a shows that a sample population containing NPT (=10) points is divided into NGS (=2) complexes. while the other (marked by •) is converging toward the global optimum.. 1993). 1994). RS MINERVE – Technical Manual Page 66/94 . NPG=m and NPT=s The SCE-UA method is explained in Figure 14 and Figure 15 for a two dimensional case (Duan et al.Chapter 4: Calibration of hydrological models Figure 13 Flow chart of the shuffled complex evolution method (from Duan et al. with V=n..2) and a local optimum located at (1.

i. The black dots (•) indicate the locations of the points in a complex before the evolution step is taken.. The probability distribution is specified such that the better points have a higher chance of being chosen to form the sub-complex than the worse points. 1994). The CCE algorithm is graphically illustrated in Figure 15. There are three types of evolution steps: reflection. contraction and mutation.e. the worst point is discarded and replaced by the new point. forms a triangle in this case) points is selected according to a prespecified probability distribution to initiate an evolution step. Figure 15b and Figure 15d illustrate the "reflection" step. Thus an evolution step is completed. the new point is generated by a "contraction" step (the new point lies half-way between the worst point and the centroid of the other points). The symbol (*) represents the new points generated by the evolution steps. A sub-complex containing NPS (=3. Since the reflected point has a lower criterion value than the worst point. which is implemented by reflecting the worst point in a sub-complex through the centroid of the other points.Chapter 4: Calibration of hydrological models Figure 14 Illustration of the shuffled complex evolution (SCE-UA) method (from Duan et al. In Figure 15c. after rejecting a reflection step for not improving the criterion value. RS MINERVE – Technical Manual Page 67/94 . Figure 15a.

Finally. i. outside of the feasible parameter space. This is realized after a reflection step is attempted.. but results in a wrong point. 1994). the Figure 15f shows the final complex after NSPL (=5) evolution steps.e.Chapter 4: Calibration of hydrological models In Figure 15e. Another scenario in which a mutation step is taken is when both the reflection step and the contraction step do not improve the criterion value. RS MINERVE – Technical Manual Page 68/94 . Figure 15 Illustration of the evolution steps taken by each complex (from Duan et al. a "mutation" step is taken by random selection of a point in the feasible parameter space to replace the wrong point of the sub-complex.

Then.4 Uniform Adaptive Monte Carlo Model architecture The Uniform Adaptive Monte Carlo (UAMC) algorithm is based on the Monte Carlo experiments that rely on repeated random sampling to obtain simulation results (Gilks et al..001 Random 4. is equal to NPG. depending on a SEED value.. which is assumed equal to MINGS according to the Duan investigation (Duan et al. The number of shuffling loops (KSTOP) in which the criterion value must change by a fixed percentage (PCENTO) before optimization is finished. 1996. NSPL. Liu. The number of complexes is defined as NGS.Chapter 4: Calibration of hydrological models Algorithm parameters Different parameters of the SCE-UA have to be defined by the user (Table 23). Table 23 Parameters of the SCE-UA algorithm Object SCE-UA Name Units Description Default Value MAXM - Maximum number of iterations NGS - Number of complexes KSTOP - Number of shuffling loops 10 PCENTO - Criterion value on shuffling loops 0. Afterwards. An initial set of Nopt parameters is given by the user or is assumed as random depending on the user’s needs and the used hydrological models. The process is repeated until the optimization converges to the best set of parameters (Figure 16). the solution space is adjusted and a new group of simulations starts. 2004). The number of points NPG in each complex corresponds to 2·Nopt+1 and the number of points NPS in each sub-complex to Nopt+1 (It has to be noted that each point corresponds to a set of parameters). RS MINERVE – Technical Manual Page 69/94 . The number of evolution steps allowed for each complex before complex shuffling. the total number of points NPT in the entire sample population is NGS·NPG. Three different convergence criteria are defined by the user:    The maximum number of function evaluations (or iterations) MAXN. but has been modified in order to iteratively adjust the solution space. and other parameters are directly calculated by the process. as presented hereafter. The algorithm randomly launches a collection of simulations (block) and finds the better results in the solution space.1 PEPS - Convergence parameter SEED - Seed value 10000 3 0. The other NPT-1 points (or parameters sets) are randomly created by the algorithm. 2001) . The PEPS parameter which provides a flag indicating whether parameter convergence is reached (It compares the value of PEPS with the normalized geometric mean of parameter ranges).

1 0. A number of iterations ITGR per group is defined for the optimization. This solution space takes into account the minimum and the maximum values of each parameter providing the best values and adds an additional range COEFRANG. Finally. as presented hereafter. a number NUMBEST of best values is applied for calculating the solutions space range for the next group of iterations. the optimization finishes when the convergence criterion (defined as ERR) is achieved. Table 24 Parameters of the UAMC algorithm Object UAMC Name Units MAXN - Maximum number of iterations 2000 ITGR - Number of iterations per group 100 NUMBEST - Number of best values taken into account for the next group calculation COEFRANG - Additional range coefficient ERR - Error difference until convergence SEED - Seed value RS MINERVE – Technical Manual Description Default Value 5 0.001 Random Page 70/94 .Chapter 4: Calibration of hydrological models Algorithm parameters Different parameters of the UAMC algorithm have to be defined by the user (Table 24). Random values of parameters are used for each iteration of the group based on a SEED value. or when the maximum number of iterations MAXN is attained. Once the first group of iterations is finished.

The advantage of this subroutine calculation lies in the speed to obtain near optimal values. Scanning of the space of possible solutions is performed by the Latin Hypercube. 1979) and the Rosenbrock algorithm (Rosenbrock. An important advantage of this method is that the dimension of the problem is defined by the division of the latin hypercube and not by the number of parameters. The best results from samples become the starting points required for Rosenbrock algorithm. called hereafter CLHR. thus reducing the number of evaluations of the objective function. 1960). with more homogeneous samples achieved with fewer samples. This combined algorithm can discretize a wide domain and then narrow your search to smaller sectors (Figure 17). generates a powerful tool for optimization of complex problems.Chapter 4: Calibration of hydrological models Start Initial range per parameter SEED value ITGR iterations results Stop Yes Convergence (ERR) achieved or MAXN attained? No New Range per parameter depending on NUMBEST and COEFRANG: [New_min=minnumbest-(maxnumbest-minnumbest)*COEFRANG.. This algorithm allows pseudo-statistical sampling conditioned by the previous calculated solutions. RS MINERVE – Technical Manual Page 71/94 . adjusting axis changes based on the direction of maximum enhancement. This algorithm is based on a gradient search. New_max=maxnumbest + (maxnumbest-minnumbest)*COEFRANG] Figure 16 Flow chart of the UAMC algorithm 4.5 Coupled Latin Hypercube and Rosenbrock Model architecture The Coupled Latin Hypercube (McKay et al. The Latin hypercube is an evolution of Monte Carlo method.

Step 2 Rank points: The results obtained in the first step are ordered (Figure 21). RS MINERVE – Technical Manual Page 72/94 .Chapter 4: Calibration of hydrological models Figure 17 Illustration of the operation of the CLHR algorithm. uncertainty is reduced within the domain of search. The best results from the Latin hypercube algorithm will serve as starting points to launch the Rosenbrock algorithm. If the size of the hypercube is greater. A general description of the steps of the CLHR method is given below: Step 1 Generate sample: Generation and evaluation of a pseudorandom sample by a Latin hypercube within the feasible parameter space.

it should have obtained worse results in all directions of search and at least an improvement in one of these direccions (remember each direction has 2 ways). Step 4 Axes change: Axes are changed to orient the Cartesian axes to the direction of maximum improvement. For applying this change of axes. This subroutine searches around the starting point the values that improve the objective function. Depending on whether the objective function improves or worsens.Chapter 4: Calibration of hydrological models Figure 18 Illustration of the operation of the Latin Hypercube algorithm. Step 5 The best result of all Rosenbrock releases is stored. the parameters values are changed to advance or backward. RS MINERVE – Technical Manual Page 73/94 . Step 3 Launch of Rosenbrock: Rosenbrock algorithm starts at least once.

The BETA coefficient represents the movement if a worse result is obtained. as presented in eq. The RLAUNCHES (1≥RLAUNCHES≤DivLH) best results from the Latin Hypercube algorithm are used as starting points for the Rosenbrock algorithm.The SEED is responsible for generating the randomness of the sample.1. The STEPROS parameter indicates the subdivisions for each parameter’s range. or when the maximum number of iterations MAXN is attained. 𝑀𝑎𝑥𝑃𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟𝑉𝑎𝑙𝑢𝑒𝑖 − 𝑀𝑖𝑛𝑃𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟𝑉𝑎𝑙𝑢𝑒𝑖 CLHR.Chapter 4: Calibration of hydrological models Figure 19 Illustration of the operation of the Rosenbrock algorithm. It is used to calculate value variations in each of the parameters (∆i) to be studied.2 = ∆𝑖 𝑆𝑇𝐸𝑃𝑅𝑂𝑆 The optimization finishes when the convergence criterion (defined as ERR) is achieved. The ALPHA coefficient represents the increment in the direction of search if the objective function improves. CLHR. RS MINERVE – Technical Manual Page 74/94 . Algorithm parameters Number of tests with the latin hypercube algorithm is equal to the parameter DivLH (≥ 2).

Chapter 4: Calibration of hydrological models Table 25 Parameters of the CLHR algorithm Object CLHR Name Units MAXN - Maximum number of iterations DivLH - Latin hypercube division SEED - Seed value RLAUNCHE S - Rosenbrock algorithm launches 2 ALPHA - Advance coefficient 3 BETA - Backward coefficient STEPROS - Parameter range subdivision ERR - Convergence parameter RS MINERVE – Technical Manual Description Default Value 2000 50 Random -0.001 Page 75/94 .5 40 0.

1 Model files Both Save and Export generate files of the same format (.rsm format as well as different other formats for results of configuration data. 5.rsm). whereas with Export. as presented in the example of the Figure 20. the entire model is saved.2 Parameters and initial conditions files Parameters (or initial conditions) of the model can be saved as:   Excel files (. Text files (. With Save. 5.Chapter 5: Files formats Chapter 5: Files formats RS MINERVE generates model files saved in . only the objects (submodels included) contained in the active hierarchical level are exported.xlsx). Each hydrological object is presented per line with its parameters or initial conditions Figure 20 Example of the parameters of the GSM hydrological models in an excel file RS MINERVE – Technical Manual Page 76/94 . Each worksheet presents the parameters (or initial conditions) of one hydrological object.txt).

the number of selected series (second line).chk file 5.dbb and .csv). RS MINERVE – Technical Manual Page 77/94 .4 Exported results files The results in the model selection can be also achieved from the “Selection and plots”.dsb and . for use’s working with Aquatool DMA from UPV.Chapter 5: Files formats 5. Figure 21 Example of selected nodes saved in a .dbt and .chk file is saved. The format of this file is presented in Figure 21. Dataset files in text format (. with information such as the selected objects. with the name of the .dsx).dsx) or in binary format (. or the objective function to take into account. Aquatool input file (. the name submodel (if it exists). 5. parameters to calibrate. the name of the selected object and its chosen series.rsm model.apo). a . It contains:    the name of the selection (given by the user in the “Selection and plots”) in first line. Databases contains two different files   Database files in text format (. Some file formats are possible:    Excel files (.dbx) or in binary format (. Once the selection is exported.xlsx).5 Calibration configuration files The full configuration of the calibration. is saved in a . CSV file (.3 Objects selection files A selection of the objects data can be saved with the help of the “Selection and plots”. all selections.dst and .6 Database files The format of the database is explained in detail in chapter 6.dbx). 5.cbt file.

…) 1 Contains one or more stations Station Information about the station (name and coordinates) Contains one or more sensors Sensor Description of the sensor (name. accessible from the RS MINERVE window.1 The RS Database Viewer The RS Database Viewer window appears when a database is created ( for edition ( Open. The RS Database Viewer. is used to create or edit the database linked to the active model. is organized in five hierarchical levels:  -     - Database Description of the database Complete set of data Contains one or more groups Group Separation based on category of data (Measures. 6. 3 Definition and use of Groups and Datasets can also be done in a different way by the user. Simulations. Forecasts. units and data) Contains one data series The user can create a complete database or different datasets to add to an existing database. New) or opened The database structure. presented in Table 26. then Edit). Flow data. RS MINERVE – Technical Manual Page 78/94 .Chapter 6: Database formats Chapter 6: Database formats The different input data as well as exported results are managed within a database.) 3 Contains one or more datasets Dataset Set of data of common type (Meteo data.

1884 m3/s 3.12.2485 m3/s Station Sensor 4 Data can be interpolated in three different ways: linear.2013 14:30 UTC+01:00 (Full path of file) 21 Ko 161 Ko Description Name of the group Measure Type of data (Inputs. With constant before. RS MINERVE – Technical Manual Page 79/94 .2012 13:32 04. whereas the constant after method applies the value for the following time interval.12. the value for a given time is used as a constant for the preceding time interval.Chapter 6: Database formats Table 26 Structure of the database interface. with editable information in blue Node Database Name Description Example Description Date creation Date modification Time zone Filename Xml file Data file Name of the database Date of creation of database Date of last modification Tine scale of the data File name and path XML file size Data file size Database_tutorial 14.12.2012 00:00 14.12.01. Outputs or Inputs None) Date of last dataset capture date 18/12/2012 10:54 in the group Group Category Date capture Dataset Description Dataset creation Dataset reception Data modification Data capture Time zone Filename Xml file Data file Name of the dataset Dataset creation date Date of receipt of data Date of last modification First date of the data (editable) Tine scale of the data File name and path XML file size Data file size Meteo_data 14. constant before and constant after.12.2012 03:00 14.2012 00:00 UTC+01:00 (Full path of file) 21 Ko 161 Ko Name X Y Z Name of the station Coordinate X in meter Coordinate Y in meter Altitude in meter as sea level Station_meteo_1 650000 130000 1957 Description Category Unit Interpolation 4 Initial date Final date Min value Average value Max value Name of the variable Variable type Variable unit Data assimilation method The first date of the time series The last date of the time series Minimum value of the series Average value of the series Maximum value of the series Precipitation Precipitation MillimetersPerHour Linear 01/01/2011 00:00:00 31/12/2013 00:00:00 0 m3/s 1.2012 03:00 14.

2 Database and dataset formats Databases and datasets also correspond to different saving formats:  Database files ([. stations and sensors) A file containing the data series. This file can be in text (*. each data series should be composed of: - - 5 A header containing four fields separated by a backslash character: Group\Dataset\Station\Sensor.dbx) describing the structure of the database (groups.dbt and .dbb and . datasets.dbb) format.dst and .dbb) will be empty. Two columns (date and data) separated by a tabulation character: If the database is not containing sensor.dsx] in text format or [.Chapter 6: Database formats 6. Each field name must be identical to the name found in the corresponding <Name> node in the xml file (*. Note that in the text file (*. if the database is not empty and contains one or more sensors (then at least a dataset with at least a station)5.dbt) or binary (*.dbx) (Figure 22). Dataset files ([.dbt or *.dbx] in text format or [. This last format allows to reduce the file size.  Database files The database files includes two different files: - A xml file (*.dbx] in binary format) can be created or saved from the complete database (database level).dbt).dsx] in binary format) can also be created or exported from the RS Database (dataset level). the data file (*.dsb and . RS MINERVE – Technical Manual Page 80/94 .

3 … Figure 22 Format of the two database files (in text format for the data file) RS MINERVE – Technical Manual Page 81/94 .dbt file: Group1\Dataset1\EVO\P 31.03.Chapter 6: Database formats *.2014 15:00 57.dbx file: … <Group> … <Name>Group1</Name> … <Dataset> … <Name>Dataset1</Name> … <Station> … <Name>EVO</Name> … <Sensor> … <Name>P</Name> … </Sensor> … </Station> … </Dataset> … </Group> … *.

xml file < DataBase> <DateCreated>2013 -02-22T11:26:47.5426565+01:00</DateCreated> <Version>4.4256221+01:00</DateCreated > <DateRecept>2013 -02-22T11:27:04.22T11:27:04.001</Version> <Key>4511a589 -698d-462f -8980-959d29110b90</Key> <ID>0545ddf0 -5c60 -4bd4 -a845 -b346f0598b72</ID> <LastComputer>N29094</LastComputer> <LastUser /> <Name>Databas e _tutorial</ Na me > <LastDateModified>2013-02-22T11:27:16. </Station> </Stations> </DataSet> </DataSets> <Category>None</Category > </Group> </Groups> <TimeOffset>0</Tim e Offset> <Inputs /> <Outputs /> </DataBase> General inf ormation about the database Groups section General inf ormation about the group Stations section First station Sensors sub-section First sensor Second sensor Coordinates of the f irst station Figure 23 Example of the database xml file RS MINERVE – Technical Manual Page 82/94 .02..4256221+01:00</DateCapture> <Stations> <Station> <Key /> <Name>EVO</ Na me > <Sensors> <Sensor> <Key /> <Type>Inputs</Type> <Name>P</Nam e> <Category>Precipitation </Category > <Unit>MillimetersPerHour</ Unit> <InterpolationMode>Linear</InterpolationMod e> </Sensor> <Sensor> <Key /> <Type>Inputs</Type> <Name>T</N am e> <Category>Temperature </Cat egory > <Unit>DegreeCelsius </Unit > <InterpolationMode>Linear</InterpolationMod e> </Sensor> </Sensors> <X>605415 </X> <Y>106740 </Y> <Z>1825</Z> </Station> ….001</Version> <Key /> <ID /> <LastComputer /> <LastUser /> <Name>Dataset1 </ Nam e > <DateCreated>2013 ..0912894+01:00</LastDateModified > <Groups> <Group> <Key>ad506514 -af29 .Chapter 6: Database formats An example of the xml file and of the text file is showed respectively in Figure 23 and Figure 24.4356227+01:00</DateRecept> <LastDateModified >2013 -02-22T11:27:16.467c-9eea-4fa788c038c8</Key> <Name>Group1 </Name > <DataSets> <DataSet> <Version>4.0912894+01:00</LastDateModified > <DateCapture>2013-02-22T11:27:04. ….

Chapter 6: Database formats

txt file
Group1 \Dataset1 \EVO\P
27.11.2012 00:00:00
27.11.2012 01:00:00
27.11.2012 02:00:00
27.11.2012 03:00:00
27.11.2012 04:00:00
27.11.2012 05:00:00
27.11.2012 06:00:00
Group1 \Dataset1 \EVO\T
27.11.2012 00:00:00
27.11.2012 01:00:00
27.11.2012 02:00:00
27.11.2012 03:00:00
27.11.2012 04:00:00
27.11.2012 05:00:00
27.11.2012 06:00:00

Group\Dataset\Station\Sensor
0
0
0
0
0
0
0

Date (JJ.MM.YYYY hh:mm:ss)
& value

2.5
5.7
4
3.4
4.1
0.6
0.6

Figure 24 Example of the database data file in text format

Dataset files
The dataset files includes also two different files (Figure 26 and Figure 27):
-

A xml file (*.dsx) describing the structure of the dataset (stations and sensors)
A file containing the data series. This file can be in text (*.dst) or in binary (*.dsb)
format. This last format allows to reduce the file size.

Note that in the text file (*.dst), if the dataset is not empty and contains one or more sensors
(then at least a dataset with at least a station)6, each data series should be composed of:
-

-

6

A header containing two fields separated by a backslash character: Station\Sensor.
Each field name must be identical to the name found in the corresponding <Name>
node in the xml file (*.dsx) (Figure 23);
Then two columns (date and data) separated by a tabulation character.

If the dataset is not containing sensor, the data file (*.dst or *.dsb) will be empty.

RS MINERVE – Technical Manual

Page 83/94

Chapter 6: Database formats

*.dbx file

<Station>

<Name>EVO</Name>

<Sensor>

<Name>P</Name>

</Sensor>

</Station>

*.dst file
EVO\P
31.03.2014 15:00

57.3

Figure 25 Format of the two dataset files (in text format for the data file)

An example of the xml and text files is showed respectively in Figure 26 and Figure 27.

RS MINERVE – Technical Manual

Page 84/94

Chapter 6: Database formats

xml file
<DataSet >
<Version>4.001</Version>
<Key></Key>
<ID></ID>
<LastComputer/>
<LastUser/>
<Name>Dataset</N ame >
<DateCreated>2014 -04-15T11:15:22.3963143+02:00</DateCreated>
<DateRecept>2014 - 04- 15T11:15:22.3963143+02:00</DateRecept>
<LastDateModified>2014 -04-15T11:28:49.904057+02:00</LastDateModified>
<DateCapture >2014 -04-15T11:15:22.3973144+01:00</DateCapture >
<Stations>
<Station>
<Key/>
<Name>EVO</N ame >
<Sensors>
<Sensor>
<Key/>
<Type>Inputs</Type>
<Name>P</Name>
<Category>Precipitation</Category>
<Unit>MillimetersPerHour</Unit>
<InterpolationMode >Linear</InterpolationMode >
</Sensor>
<Sensor>
<Key/>
<Type>Inputs</Type>
<Name>T</Name >
<Category>Temperature </Category>
<Unit>DegreeCelsius</Unit>
<InterpolationMode >Linear</InterpolationMode >
</Sensor>
</Sensors>
<X>605415 </X>
<Y>106740 </Y>
<Z>1825</Z>
</Station>
<Station>
<Key/>
<Name>AIG</N ame>
<Sensors>
<Sensor>
<Key/>
<Type>Inputs</Type>
<Name>P</Name>
<Category>Precipitation</Category>
<Unit>MillimetersPerHour</Unit>
<InterpolationMode >Linear</InterpolationMode >
</Sensor>
<Sensor>
<Key/>
<Type>Inputs</Type>
<Name>T</Name >
<Category>Temperature</Category>
<Unit>DegreeCelsius</Unit>
<InterpolationMode >Linear</InterpolationMode >
</Sensor>
</Sensors>
<X>560401 </X>
<Y>130713 </Y>
<Z>381</Z>
</Station>
</Stations>
</DataSet>

General
inf ormation
about the
dataset

Stations section
First station
Sensors sub-section

First sensor

Second sensor

Coordinates of the f irst station
Second station

Figure 26 Example of the dataset xml file

RS MINERVE – Technical Manual

Page 85/94

11.2012 00:00:00 27.2012 06:00 AIG\P 27.6 0.2012 03:00 27. all the possible nodes are very useful but not all required.dsx).11.11.11.11.11.11.11.2012 06:00 EVO\T 27.3 8.2012 04:00 27.2012 03:00:00 27.11.2012 02:00 27.11.11. additional information for the automatic creation is presented hereafter.11.11.2012 05:00 27. since some of them are also necessary for operational hydro-meteorological systems. RS MINERVE – Technical Manual Page 86/94 .2012 01:00 27.3 Figure 27 Example of the dataset data file in text format 6. it can be easily created based on information provided in Chapter 6.6 8.MM.7 4 3. Regarding the .2012 05:00:00 27.11.9 8.2012 06:00:00 AIG\T 27.11.11.2012 02:00 27.2012 04:00 27.2012 00:00 27.3 7 6.2012 04:00:00 27.11. since it can be used when datasets with same “Key” (as well as same filename and structure) are overwritten at each new forecast.2012 05:00:00 27.2012 01:00 27.1 0.11.dsx).11.4 0.11.11.dbx or *. “ID” node allows to RS Database knowing if the dataset has been updated or not.11.2012 04:00:00 27.2012 02:00:00 27. Concerning the text file (respectively *.6 0 0 0 0 0 0.2012 01:00:00 27.4 4.11.dbx) and a dataset xml file (*.2012 00:00 27. “ID” node is only useful for operational system.11.2012 03:00 27. The information visible in RS Database is also presented in last three columns. The presented tables takes into account an example with adataset “Measure” with a station “Aigle” and a sensor of observed precipitation “P”.5 5.xml file (*.2012 01:00:00 27. “Key” node generally allows distinguishing different databases (or datasets) with different name and structure.2012 05:00 27.2012 02:00:00 27.dst). To create an xml file.11.2012 03:00:00 27. The Table 27 and Table 28 describe all the possible nodes respectively represented in a database (*.11.3 7.YYYY hh:mm:ss) & value 2.2012 00:00:00 27.3 Automatic creation of databases or datasets An automatic creation of a database or a dataset is possible. Then.11.2012 06:00:00 0 0 0 0 0 0 0 Date (JJ.7 7.2.dbt or *.Chapter 6: Database formats txt file Station\Sensor EVO\P 27.11.

001</Version> <Key>a3e0e759-6f10-4a93-8aa1cb213ac767f1</Key> <ID>a5a3c179-5eb0-47ab-8d9cf820e741977c</ID> <LastComputer>vmcrealp <\LastComputer> <LastUser>crealp</LastUser> <Name>201404090700_Measure</Name> <DateCreated>2014-0410T10:00:12+02:00</DateCreated> <DateRecept>2014-0410T15:29:36.2862121+02:00</LastDateModified> <TimeOffset>36000000000</TimeOffset> <Groups> <Group> <Key>7f63d446-5eb7-4d75-8d2ae84c79a6bf89</Key> <Name>Measure</Name> <DataSets> <DataSet> <Version>4.1625449+01:00</DateCreated> <Version>4.0000000+02:00</DateCapture> <Stations> <Station> X X X Key Example RS MINERVE – Technical Manual Dataset description Starts the stations node Starts the description of a station To distinguish stations in the database Station description Starts the sensors node Starts the description of a sensor To distinguish sensors in the database Type of data Sensor description Sensor description Sensor description Sensor description Finish the description of a sensor Finish the sensors node Station description Station description Station description Finish the description of a station Finish the stations node Finish the description of a dataset Finish the datasets node Group description Finish the description of a group Finish the groups node Database description Finish the database node X X X X X X X X X X X X X X X X X X X X X X X X X X X Page 87/94 .Chapter 6: Database formats Table 27 Description of all the possible nodes in a database xml file Node DataBase Required Suggested X DateCreated X Version X Key X ID X LastComputer LastUser Name X X X LastDateModified X TimeOffset Groups Group X X X Key Name DataSets DataSet Version X X X X X Key X ID X LastComputer LastUser Name X X X DateCreated X DateRecept X LastDateModified X DateCapture X Stations Station X X Key Name Sensors Sensor X Type Name Category Unit X X X InterpolationMode X Sensor Sensors X Y Z Station Stations DataSet DataSets Category Group Groups TimeOffset DataBase X X X X X X X X X Database description X To verify RS Database version X To distinguish different databases X To distinguish databases with identical Key Computer origin User Database description X X X X Database description Database description Starts the groups node Starts the description of a group To distinguish groups in the database Group description Starts the datasets node Starts the description of a dataset To verify RS Database version X X X X X X X X X To distinguish different datasets To distinguish datasets with identical Key Computer origin User Dataset description X X X X X Dataset description X <LastDateModified>2014-04Dataset description 10T15:29:36.0541902+02:00</DateRecept> <DateCapture>2014-0409T07:00:00.001</Version> <Key>2234d822-7fb7-4d77-afb5564f029f4c66</Key> <ID>d11fa701-0e05-4833-847e4385a8b808f1<\ID> <LastComputer>D29493</LastComputer> <LastUser>D29493</LastUser> <Name>Database</Name> <LastDateModified>2014-0410T15:31:46.0531901+02:00</LastDateModified> X <Key>AIG</Key> <Key>AIG\P</Key> X <Type>Inputs</Type> <Name>P</Name> <Category>Precipitation</Category> <Unit>MillimetersPerHour</Unit> <InterpolationMode>ConstantBefore</Interpola tionMode> </Sensor> </Sensors> <X>560401</X> <Y>130713</Y> <Z>381</Z> </Station> </Stations> </DataSet> </DataSets> <Category>Outputs</Category> </Group> </Groups> <TimeOffset>36000000000</TimeOffset> </Database> X Starts the database node Visible Non visible Editable Non editable X X X X Purpose Dataset description <Name>Aigle</Name> <Sensors> <Sensor> X X X <Database> <DateCreated>2014-0205T17:10:10.

001</Version> <Key>a3e0e759-6f10-4a93-8aa1cb213ac767f1</Key> <ID>a5a3c179-5eb0-47ab-8d9cf820e741977c</ID> <LastComputer>vmcrealp <\LastComputer> <LastUser>crealp</LastUser> <Name>201404090700_Measure</Name> <DateCreated>2014-0410T10:00:12+02:00</DateCreated> <DateRecept>2014-0410T15:29:36.0000000+02:00</DateCapture> <Stations> <Station> <Key>AIG</Key> <Name>Aigle</Name> <Sensors> <Sensor> Key X <Key>AIG\P</Key> Type Name Category Unit X X X X InterpolationMode X Sensor Sensors X Y Z Station Stations DataSet DataSets X X X X X X X X X <Type>Inputs</Type> <Name>P</Name> <Category>Precipitation</Category> <Unit>MillimetersPerHour</Unit> <InterpolationMode>ConstantBefore</Interpol ationMode> </Sensor> </Sensors> <X>560401</X> <Y>130713</Y> <Z>381</Z> </Station> </Stations> </DataSet> </DataSets> RS MINERVE – Technical Manual Dataset description Starts the stations node Starts the description of a station To distinguish stations in the Station description Starts the sensors node Starts the description of a sensor To distinguish sensors in the database Type of data Sensor description Sensor description Sensor description Sensor description Finish the description of a sensor Finish the sensors node Station description Station description Station description Finish the description of a station Finish the stations node Finish the description of a dataset Finish the datasets node X X X X X X X X X X X X X X X X X X X X X X Page 88/94 .0541902+02:00</DateRecept> Purpose Starts the datasets node Starts the description of a dataset To verify RS Database version Visible Non visible Editable Non editable X X X To distinguish different datasets To distinguish datasets with identical Key Computer origin User Dataset description X X X X X Dataset description X Dataset description X <LastDateModified>2014-04Dataset description 10T15:29:36.Chapter 6: Database formats Table 28 Description of all the possible nodes in a dataset xml file Node DataSets DataSet Version Required Suggested X X X Key X ID X LastComputer LastUser Name X X X DateCreated X DateRecept X LastDateModified X DateCapture X Stations Station Key Name Sensors Sensor X X X X X X Example <DataSets> <DataSet> <Version>4.0531901+02:00</LastDateModified> X <DateCapture>2014-0409T07:00:00.

Bitmap RS MINERVE – Technical Manual Extensions *. *. *.bmp Page 89/94 . *. *. *.wmf *. *. *.emf.ico.shp *.Chapter 7: GIS formats Chapter 7: GIS formats RS MINERVE includes a GIS module.wmf *.bmp. *. *. *. *.gif.jpg.shp *. *.bgd *.png.bmp. *. *.bmp. *.exf.png.jpg. 7. *. *.gif. *. Table 29 Formats and extensions supported by RS GIS Formats Vector Raster Image DotSpatial.exf.emf.tif.Shapefile DotSpatial – Binary file DotNet Image .bmp.1 Model files Layers with formats presented in Table 29 can be imported in RS GIS.ico.Image DotNet Image .tif.bgd *.

air. Norrköping. B. Ferral. J. F. (1980) Polycopié « Hydraulique urbaine ». Norrköping.Distributed Hydrological Modelling. Ph.F. (Ed.D.A. (1998). Cunge. C. p 204. R. Ajami. H. (1939). Statistics for Spatial Data.A. N. Laboratoire de Constructions Hydrauliques. 133-156.. M. Mathematical. Experiments with Fluid Friction in Roughened Pipes. P. and White. National Weather Service and State of California. (1995).. Colebrook. Wiley RS MINERVE – Technical Manual Page 90/94 . Institut National Polytechnique de Grenoble. Turbulent flow in pipes. (1991). Burnash. 4.-S. 451-476. Cressie. S. 311-366. Application of a Conceptual Hydrologic Model in Teaching Hydrologic Processes. Thesis. International Journal of Engineering Education 26(4). and Rosbjerg. J. The NWS River Forecast System . R. Series A. US Department of Commerce. Habib. J. Niggli M. Department of Water Resources. Journal of Hydrology 298. Luft – eau. E. (2007). (1973). A. and Refsgaard. C. The HBV model . R. C. 321 pp.). Heft 9/10. with particular reference to the transition region between smooth and rough pipe laws. In: Singh. R. 7. Application au bassin versant supérieur du Rhône.its structure and applications. Physical and Engineering Sciences 161. Development and application of a conceptual runoff model for Scandinavian catchments. J. A.catchment modeling. Ecole Polytechnique Fédérale de Lausanne. (1996). (1969). Journal of the Institution of Civil Engineers 11. Concepts méthodologiques pour le calcul des crues. Energie. Wagener. 367-381. (1992). Au sujet d’une méthode de calcul de propagation des crues (Méthode Muskingum). T. 4-5. 223-231. Springer Science & Business Media. V. C. C.. N. H. No. 1973. AghaKouchak.. M. A generalized streamflow simulation system – Conceptual modelling for digital computers. No. Wasser. Calibration of a semidistributed hydrologic model for streamflow estimation along a river system. (1976). (1993). and Musy A. K. S. Cunge. 112-135. R...Bibliography Bibliography Abbott. (2004). L. J. Bergström. and McGuire. C. Burnash. Bergström. Parameter estimation in distributed hydrological modelling: comparison of global and local optimisation techniques. S.A. (2010). Proceedings of the Royal Society of London. énergie.-L. Polycopié Simulation des écoulements non-permanents dans les rivières et canaux. Journal of Hydraulic Research 7. Gupta. Computer Models of Watershed Hydrology. SMHI Reports RHO No. Colebrook. Nordic Hydrology. D. Consuegra D. Ecole Nationale Supérieure d’Hydraulique et de Mécanique de Grenoble.. Madsen. Blasone.C. (1937). 963-973. Vol 38. SMHI Reports RH. 205-230.. Boillat J. and Sorooshian.

Duan. 930-943. G. Recherches Expérimentales Relatives au Mouvement de l’Eau dans les Tuyaux [Experimental Research on the Movement of Water in Pipes].. J. Ecole Polytechnique Fédérale de Lausanne.. Optimal use of SCE-UA global optimization method for calibrating watershed models. F. H. 268 p. S. J. and Berlamont. (1994). Chapman and Hall. D. 573-587. 19 (3). Duan. Adding sediment transport to the integrated hydrology model (InHM): Development and testing. J. Routing System II. Sorooshian. 28. 512p. Environmental Modelling and Software 38. No. L. J.. García Hernández. and Sorooshian.18. (2000). Lausanne. (1996) Markov Chain Monte Carlo in Practice. J. (1993).. J. Ed. K... Propagation of uncertainty from rainfall to runoff: A case study with a stochastic rainfall generator. 265-284.. S. Switzerland.295 RS MINERVE – Technical Manual Page 91/94 . (1857). Hydrological Processes. Advances in Water Resources 29 (6). Vol. Gilks.. pp. & Marshall. Journal of Lake Sciences. Sharma. and Gupta. 158. B. Foehn. L. J. S. (2011). García Hernández. V. Duan. S. Journal of Optimization Theory and Applications. G. and Boillat.. (1992). and Provenzale. Arnold. Vanderkwaak. Vázquez. No. S. 304-314. & atlas. J. A shuffled complex evolution approach for effective and efficient global minimization. Gabellani. and Spiegelhalter. (2005).. Flow modelling in hydraulic systems. El-Nasr. Sels. O. Communication 9 du Laboratoire de Constructions Hydrauliques.. Heppner. J. Richardson.A. Christiaens. Hapuarachchi H. R. Ed.. Vol.Bibliography Darcy. Flood management in a complex river basin with a real-time decision support system based on hydrological forecasts... RS MINERVE Group. E. Application of a distributed physically-based hydrological model to a medium size catchment.A. V. K. (2000). Feyen. and Feyen. 2061-2071. A. 4. 283 . S. Vol. 10151031.. (2001).. S.. No. A. Routing System . 4. Jeremiah. L.P. J. Ferraris. Effective and Efficient Global Optimization for Conceptual Rainfall-Runoff Models. Gupta. 47-63. W. EPFL. J. RS MINERVE – User’s manual v1. C. Q. Jordan. Sisson. and Boillat.. S. V. (2012). Dubois. PhD Thesis N°5093. Ran. Schleiss. LI. Roquier. R. Q.. E. and Gupta. London.Modélisation du routage des crues dans des systèmes hydrauliques à surface libre. J. A. Feyen.-L.. and Paredes Arquiola.. K. A. Communication 32 du Laboratoire de Constructions Hydrauliques. Hydrology and Earth System Sciences 4 (1). K (2006). and Loague. (2007). Schleiss. Z. (2015). von Hardenberg. García Hernández. 501-521. (2007). Water Resources Management. Vol 12. Q. 76. Wang. Q. A. J. Boni. Efficient hydrological model parameter optimization with Sequential Monte Carlo sampling.. Paris.-L. Application of SCE-UA Method for Calibrating the Xinanjiang Watershed Model. Journal of Hydrology. Switzerland. Vol. J. Mallet-Bachelier. Modelling the hydrology of a catchment using a distributed and a semi-distributed model. Dubois. A. A. 3. Sorooshian. J. Advances in Water Resources 30. Lausanne.

D. In Press.G. Justi Luvizotto.. R. (1965). (2005).L. E D. 1. Krause. 308-313. Ponnambalam. and Water Resources Engineers. Proceedings of the Fourth International Conference on Natural Computation. Comparison of several heuristic approaches to calibration of WATCLASS hydrologic model. Kang. F. Veith. and Bäse. and Coelho. T. A Pareto Strength SCE-UA Algorithm for Reservoir Optimization Operation. 1202-1205. McGuinness. V. Metcalf and Eddy. (2011). 239–245. J. and Soulis. T.S.S. Washington D. Muttil.N. Computer Journal. D. Cheng..amc. 71 pp. Information Sciences. (1979). Superior exploration-exploitation balance in shuffled complex evolution. J. V. Beckman. Journal of Hydraulic Engineering. Technometrics 21 (2). U. 11024DOC07/71 (NTIS PB-203289).Bibliography Jordan. No. Lausanne. Réflexion sur la mise au point d’un modèle pluie-débit simplifié sur plusieurs bassins versants représentatifs et expérimentaux. Journal of Hydrology 10. Rosenbrock artificial bee colony algorithm for accurate global optimization of numerical functions.. Van Liew. River flow forecasting through conceptual models: part 1 .. (2008). K. R. J. A. Nash.. (2011).L. E. Kamali. Ecole Polytechnique Fédérale de Lausanne. (2001).J. ISSN 0096-3003.. (2007). DOI: 10. Modèle de prévision et de gestion des crues . P. F. Department of Agriculture. Comparison of different efficiency criteria for hydrological model assessment. 181(16). Vol.F. New-York. (1970). and Liong. M. U. (1993). Nelder. M. University of Florida. A Simplex Method for function minimization. and Mead. No. CEMAGREF Antony. 20460. J. 406-412. Vol. Vol.D.L. Model evalutation guidelines for systematic quantification of accuracy in watershed simulations.. d. 40–46. Li. Harmel. R. Boyle. F. Water Resources Management 7 (1). Technical Bulletin 1452.2010. H. M. G. 7. American Society of Agricultural and Biological Engineers 50(3).. S.-Y. (2004). 89-97.12. DC.optimisation des opérations des aménagements hydroélectriques à accumulation pour la réduction des débits de crue. (1971). C.. C. and Lin. Bingner. Washington. P. A comparison of lysimeter-derived potential evapotranspiration with computed values. Corrected Proof. EPA. A. 4.-T. The performance of some real-time statistical flood forecasting models seen through multicriteria analysis. EPFL. 130. N.. Inc. Z. 282-290. No. L... F. E. Michel. Advances in Geosciences 5. 343p. Arnold. J. and Ma. (1972). 3508-3531. D. Moriasi. J. and Edijatno (1988). R. J. Inc. I. 38.C. and Conover. Thesis Report N°3711. C.. Storm Water Management Model. Final Report. Masmoudi. and Habaieb. 57-67..1016/j. Lin. J. A hybrid shuffled complex evolution approach based on differential evolution for unconstrained optimization. Springer Verlag. S. (2013). M. and Bordne. W. Available online 21 December 2010.A discussion of principles. A comparison of three methods for selecting values of input variables in the analysis of output from a computer code. McKay. L. S. Mariani. W. Vol. Applied Mathematics and Computation. Canadian Water Resources Journal. Guerra. Liu.J. and Sutcliffe. Monte Carlo Strategies in Scientic Computing. RS MINERVE – Technical Manual Page 92/94 . (2007).064. 885-900. 12. Agricultural Research Service..-Y. ISBN: 978-0-7695-3304-9.

The Computer Journal 3. Nguyen. Nyunt. Henry Darcy (1803–1858): Immortalised by his scientific legacy..E. D. A conceptual glacio-hydrological model for high mountainous catchments. (1961). Vol. T. Development of an integrated modeling system for improved multi-objective reservoir operation. M.V. 5-152. Hangzhou. X. Shepard. Vrugt. l'evaporation et l'ecoulement. 175–184. D. Bastidas. H.. Gupta. An automatic method for finding the greatest or least value of a function. J. T. des Eaux et des Forêts. Viviroli. (2006). C. Niggli. C. (2011). Zhejiang. Water resources research 29. A global sensitivity analysis tool for the parameters of multi-variable catchment models.. Mein. Diluzio. Brazil. L. A two-dimensional interpolation function for irregularly-spaced data. 4-18. Hapuarachchi. Perrin. 324(1-4). RS MINERVE – Technical Manual Page 93/94 . Wang. Proceedings 23rd National Conference ACM. Collischonn.. Q. Journal of Hydrology 377. T. H. L. (1955). R. and Srinivasan. Simmons. 12: 13-49. Hydrology and Earth System Sciences 15. J. and Robertson. Wang. Ann. Michel. (2000).. Turc. Schäfli. B. S. 73-117. 10-23. (1968). Ann. O.. C. Koike.A. China. Journal of Hydrology.E.R. and Andréassian.M. Sun. (2003). V.. A. H. J. Zhu. (2011). (2003).. Wu. and Sorooshian. B. D. Thèse. Agro.H. S. (2005).J. Gurtz. X.. Oudin.. 166-175. Turc..G. (2006). Mittelbach. 517–524. Hydrology and Earth System Sciences Discussions 2. Pagano.. INRA. T. Paz.C. T. Le bilan de l'eau des sols. Weingartner. Zhou... W. Effective and efficient algorithm for multiobjective optimization of hydrologic models. Tucci. L. Rosenbrock. 585–595. Frontiers of Architecture and Civil Engineering in China. Proceedings of the First International Multi-Symposiums on Computer and Computational Sciences. C. H. Agro. formule climatique simplifiée et mise à jour. Journal of Hydrology 239. 303-308. A. Saavedra. 6. 1023-1038. T. 275-289. Recherche d'un modèle d'évapotranspiration potentielle pertinent comme entrée d'un modèle pluie-débit global.. Bouten.. J. A. T. S. L... Evaluation des besoins en eau d’irrigation.. (2004).. Hingray. (2008). 1. 208-225. 1214. and Wu. Zhang. INRA.A. V. Bishop. ISBN: 0-7695-2581-4. Journal of Hydrology 404. No. Improvement of a parsimonious model for streamflow simulation. Continuous simulation for flood estimation in ungauged mesoscale catchments of Switzerland – Part II: Parameter regionalisation and flood estimation results. M. Application of SCE-UA to Optimize the Management Model of Groundwater Resources in Deep Aquifers of the Yangtze Delta.Bibliography Nóbrega.. Monthly versus daily water balance models in simulating monthly runoff. 4. Paris. C. 47-55.. W.. L. Relations entre les precipitations. (2009). A.D... T. R. Grunwald.F. T.. and Sap. Uncertainty in climate change impacts on water resources in the Rio Grande Basin. J. C. and Musy. L. Journal of Hydrology 279. van Griensven. Flood estimation using radar and raingauge data. (1960). Ecole Nationale du Génie Rural. Meixner. Keenan. R. (2010)..L. Hydrogeology Journal 16.P. and Elliott.A. L.

in particular: Stéphane Micheloud1. Aurélien Claude1 and Javier Fluixá Sanmartín1.UPV 3 HydroCosmos SA [SIGNET FIN DE DOCUMENT] RS MINERVE – Technical Manual Page 94/94 .CREALP 2 Université Polytechnique de Valence . Edgar Belda2. Néstor Lerma2.Acknowledgments Acknowledgments RS MINERVE is developed by the research center CREALP and the engineering office HydroCosmos SA. we would like to express our gratitude to the people who have personally contributed to the improvement of the program and its documentation. In addition. Alex Dionisio-Calado1. with the collaboration of two universities (Ecole Polytechnique Fédérale de Lausanne and Universitat Politècnica de València) and the Hydro10 Association. We would like to thank sincerely these organizations for their support. Nicolas Rey3. Samuel Alesina1. ______________________________________ 1 Centre de recherche sur l’environment alpin .

No precipitation values for station 'Station 1' (X m.1 . A.Appendix A: List of notifications from pre-simulation and post-simulation reports Appendix A: List of notifications from pre-simulation and post-simulation reports The list of notifications (notes. RS MINERVE – Technical Manual Missing database to get station data. Z masl).1 Any isolated objects Pre-simulation report Warning If an object is isolated and is not taken into account for calculation. but a precipitation input is not available in the database. A.and post-simulation reports are presented hereafter. A. No temperature/ETP values for station 'Station 1' (X m. A null series will be considered. but no datagroup is selected in the Data source. Warning If this object exists and a database is connected to the hydrological model. The object 'Object 1' has no links and is isolated. Y m. No datagroup is selected for station 'Station 1' (X m. No dataset is selected for station 'Station 1' (X m.2 Base objects Virtual Weather Station Pre-Simulation Report Fatal If a Virtual Station exists and is not isolated. Z masl). Fatal If this object exists and a database is connected to the hydrological model. Z masl). warnings and fatal) of the pre. but any dataset is selected for meteorological data utilisation. but a temperature / ETP input is not available. Fatal If this object exists and a database is connected to the hydrological model. but no database is connected to the model. Z masl). Y m. Fatal If this object exists and a database is connected to the hydrological model. Y m. Y m.

Snow-GSM model description Pre-simulation report Fatal If this object exists and is connected to other objects. Radius extended to satisfy No. RS MINERVE – Technical Manual The object 'SWMM 1' has no Precipitation 'i' as input. The object 'Glacier-GSM 1' has no Precipitation 'i' / Temperature 'T'/Height 'Hsnow' as input. Glacier-GSM model description Pre-simulation report Fatal If this object exists and is connected to other objects. but not all Glacier-GSM model inputs are available.Appendix A: List of notifications from pre-simulation and post-simulation reports Post-Simulation Report Warning If this object does not find data for Precipitation. Runoff (SWMM) model description Pre-simulation report Fatal If the object exists and is not isolated. but not all Glacier-GSM model inputs are available. but not all Snow-GSM model inputs are available. research radius is extended until finding one. The object 'Glacier-GSM 1' has no Precipitation 'i'/ETP 'ETP'as input. A. but has no intensity as input. Temperature or ETP in any station of the database within in the radius of research.2 . No precipitation found within search radius of virtual weather station 'Station 1'. min of stations. The object 'Snow-GSM 1' has no Precipitation 'i' / Temperature 'T' as input. Infiltration (GR3) model description Pre-simulation report Fatal If this object exists and is connected to other objects.

but not all GSM model inputs are available. Channel routing description No fatal. The object 'GSM 1' has no Precipitation 'P' / Temperature 'T' as input. but not all HBV model inputs are available. warnings or notes messages for this object. but not all SOCONT model inputs are available. The object 'HBV 1' has no Precipitation 'P'/Temperature 'T'/ETP 'ETP' as input. The object 'SAC' has no Precipitation 'P'/ETP 'ETP' as input. GR4J model description Pre-simulation report Fatal If this object exists and is not isolated. Junction No fatal. but not all SAC model inputs are available.Appendix A: List of notifications from pre-simulation and post-simulation reports GSM model description Pre-simulation report Fatal If this object exists and is not isolated. HBV model description Pre-simulation report Fatal If this object exists and is not isolated.3 . SOCONT model description Pre-simulation report Fatal If this object exists and is not isolated. The object 'SOCONT 1' has no Precipitation 'P' / Temperature 'T'/ETP 'ETP' as input. but not all HBV model inputs are available. The object 'GR4J' has no Precipitation 'P'/ETP 'ETP' as input. RS MINERVE – Technical Manual A. SAC model description Pre-simulation report Fatal If this object exists and is not isolated. warnings or notes messages for this object.

keeping the reservoir level. Missing database group to get HPP data. to maximum value of paired data HV when volume exceeds the maximum volume). V. Post-Simulation Report Warning If the volume becomes negative in a moment of the simulation.3 Structures objects Reservoir Pre-simulation report Fatal If this object is used and is not isolated. the simulation is computed (with a correct volume variation. Fatal If this object is used and is not isolated. Nevertheless. Nevertheless. Warning If the initial level of the reservoir is out of the range proposed in the H-V paired data. V. the simulation is computed (with a correct volume variation.4 . H. Missing database to get initial height of reservoirs. The object 'Reservoir 1' has no Flow 'Qe' as input. Hini of reservoir 'Reservoir 1'is out of range. in zero for negative volumes). The simulation is achieved. Warning If the volume exceeds the maximum volume proposed in the relation H-V. Hini of reservoir 'Reservoir 1'used as initial height. assuming the level equals to the maximum level of the reservoir. Note If a Reservoir object exists but no database is connected to the model. Note If the reservoir uses the initial condition from the database. but the H-V paired data is not provided. keeping the reservoir level.Appendix A: List of notifications from pre-simulation and post-simulation reports A. H. RS MINERVE – Technical Manual A. Volume of reservoir 'Reservoir 1' exceeds its capacity. Volume of reservoir 'Reservoir 1' contains negative values. Note If the reservoir uses the initial condition proposed by the user (and not data from database as initial condition). but does not have an inflow. Note If no group is selected in the Data source for reservoirs. Reservoir 'Reservoir 1' is missing H-V paired data. Initial height of reservoir 'Reservoir 1' taken from database.

The maximum level of HQ 'HQ 1' has been attained or exceeded. but the H-Q paired data is not provided. HQ 'HQ 1'is missing H-Q paired data. Missing database group to get TurbineDB data. Post-Simulation Report Warning The computed simulation runs even if the level is equal or over the maximum level proposed in the relation H-Q. Turbine Pre-simulation report Fatal If this object is used and is not isolated. Warning If this object if used and is not isolated. the outflow is kept as the maximum discharge existing in the relation H-Q. A.Appendix A: List of notifications from pre-simulation and post-simulation reports HQ Pre-simulation report Fatal If this object is used and is not isolated. but no group is selected in the Data source for TurbineDB. RS MINERVE – Technical Manual Missing database to get TurbineDB data. In that case.5 . The object 'Turbine 1' has no Flow 'Qwanted' as input. Warning If this object if used and is not isolated. TurbineDB Pre-simulation report Warning If a TurbineDB object exists and no database is connected to the model. but the TurbineDB has not a station from database to provide data series. but do not receive the Qwanted data as input. No station or sensor for TurbineDB 'TurbineDB 1'.

Consumer Pre-simulation report Warning If a source object exists and no database is connected to the model.Appendix A: List of notifications from pre-simulation and post-simulation reports Hydropower Pre-simulation report Fatal If this object if used and is not isolated. Missing database to get Hydropower data. No station or sensor for Consumer 'Consumer 1'. Diversion Pre-simulation report Fatal If this object if used and is not isolated.6 . Warning If no group is selected in the Data source for sources. Hydropower 'Hydropower 1'is missing paired data values. Default QDemand values will be used. RS MINERVE – Technical Manual Missing database to get Consumer data. Price defined from Default Price. Warning If this object is used and is not isolated. Missing database group to consumer data. Warning If no group is selected in the Data source for sources. but the Q-η paired data is not provided. Default Price value(s) will be used. A. but the consumer has not a station from database to provide data series. Warning If a source object exists and no database is connected to the model. Warning If this object is used and is not isolated. Consumption defined from Default QDemand. Diversion 'Diversion 1'is missing QUp-QDiverted paired data. Default QDemand values will be used. Default Price value(s) will be used. but the consumer has not a station from database to provide data series. No station or sensor for Hydropower 'Hydropower 1'. Missing database group to get Hydropower data. but the QUp-QDiverted paired data is not provided.

Missing database group to source data. but the source has no a station from database to provide data series. Comparator No fatal errors. Missing database to get source data.Appendix A: List of notifications from pre-simulation and post-simulation reports A. but the H-Q paired data is not provided. Warning If this object is used and is not isolated.7 . Times Series 'Time series 1'is missing paired data values. warnings or note messages for this object Group Interface No fatal errors. warnings or note messages for this object Sub-model No fatal errors. warnings or note messages for this object RS MINERVE – Technical Manual A. Source Pre-simulation report Warning If a source object exists and no database is connected to the model.4 Standard objects Time Series Pre-simulation report Fatal If this object if used and is not isolated. Warning If no group is selected in the Data source for sources. No station or sensor for source 'Source 1'.

8 .Appendix A: List of notifications from pre-simulation and post-simulation reports A. The regulation 'Regulation 1' has no initial state. warnings or notes messages for this object RS MINERVE – Technical Manual A.5 Regulation objects Regulation and States Pre-simulation report Fatal Error An initial state is required in the regulation. Sensors No fatal.

+41 (0) 27 607 11 80 Fax +41 (0) 27 607 11 94 crealp@crealp.A.Centre de recherche sur l'environnement alpin Rue de l'Industrie 45 CH-1951 Sion Tél.RS MINERVE CREALP . +41 (0) 27 764 34 20 Fax +41 (0) 27 764 34 21 info@hydrocosmos.vs. Grand-Rue 43 CH-1904 Vernayaz Tél.ch HydroCosmos S.ch .