1

CWRA/CGU 2012 Conference Banff, Alberta June 5-8, 2012

K. Abbasnezhadi

Parallel Interaction of the Hydrologic Regime and the Spatio-Temporal Land Use Patterns
K., Abbasnezhadi
Department of Civil Engineering, University of Manitoba, Winnipeg, Manitoba, Canada

Abstract:
The challenge here was to consider the effect of one factor, referred to as land use, on the hydrologic regime of a catchment, among all the independent, yet integral and dominant factors. Such a distinct approach was applied since land use transformations play an important role in defining the rate of changes streamflow has been posed to under different climatic conditions. To detect such spatial and temporal changes in different categories of land use types, it is highly convenient to employ GIS tools. The area under study was the Upper Langat River catchment, in Malaysia. Concurrent assessment of land use maps of 1997 and 2006, and hydrologic properties such as mean, and maximum monthly and annual discharges at the basin’s outlet station revealed that 86% growth of impervious areas, along with 13% decline of forest lands were resulted in a 48% raise of the mean annual stream discharge of the Upper Langat River by 2006. Furthermore, it was shown that mean annual discharges exhibited an increasing trend during the high-flow months, yet a decreasing trend during low-flow ones. In general, it was revealed that the basin’s potential to hold rainfall water rather than discharging it has been reduced. KEY WORDS Landuse, Langat, Spatio-temporal analysis, Hydorologic regime 1. INTRODUCTION

Commonly, the human alteration of Earth's terrestrial cover is expressed as Land Use and Land Cover Change (LULCC). The direct and indirect results of human activities to save Earth vital resources are the modifications in land use and land cover which dated back to early history. However, due to the fact that agriculture and deforestation have been there for thousands of years in many regions of the world, it is often challenging to distinct between the effects of human land use on the landscape and the natural landscape processes. Recent levels, scopes and magnitudes of LULCC are far higher than ever, causing exceptional changes in the ecosystems and environmental processes at local, regional and global scales (Erle & Pontius, 2007). In 1900, only a small fraction of the world’s population – a mere 15% – lived in the cities, while at the moment more than 50% do so, and the anticipations show that between 1990 and 2050, the urban population will increase to over 5 billion (Maksimovic & Tucci, 2001). In Malaysia, for example, the total number of urban dwellers has risen to 58.8% in the year 2000 (Hadi, 2000). Such uncontrolled growth of urban expansion has unfavorably distressed the Malaysian river basin ecosystems (Jahi & Hasan, 1996), especially their streamflow regulating capabilities. To determine the outcomes of land use changes on river flow in a given basin, two issues are normally come across. There is a complexity in resolving the changes in land use which calls for high resolution land use maps, as well as the constraints encountered when relating these changes to river flow. To overcome such issues, it is highly harmonizing to apply computer systems and information technology in handling geographic and spatial data. In this regard, the use of the geographic information systems (GIS) and remote sensing (RS) were found to be helpful gears to spot and analyse the spatio-temporal land use dynamics.

2

Parallel Interaction of the Hydrologic Regime and the Spatio-Temporal Land Use Patterns 2. PROPERTIES OF THE STUDY AREA

Langat River basin and more specifically the Upper Langat River subbasin have lately been highly exposed to fast urban and built-up development. The basin consists of four major subbasins: Upper Langat, Labu, Semenyih, and Lower Langat. However, this study was focused on the Upper Langat River subbasin. The Upper Langat River catchment encompasses an area of 385km2 and falls within the coordinates, 101⁰ 43’ E, 3⁰ 17’ N and 101⁰ 58’ E, 2⁰ 59’ N. The basin can be categorized into 6 subbasins as shown in Figure 1: Langat Dam, Sungai Lui, Langat session 1, Langat Session 2, Sungai Tekali, and Downstream Kajang from where the discharge data used in this study were taken. Towards the south west, the area is low in elevation (<40m) where the urbanization growth is mainly found. Towards the north east, there are some hilly areas surrounded by dense tropical forest, the highest point is about 1400 m. The highest amount of rainfall occurs in April and November with amounts over 250 mm, while the lowest occurs in June, with an average of 100mm (Noorazuan et al., 2003).

Langat Dam

Langat Session 1 Sungai Lui Langat Session 2 Sungai Tekali

Downstream Kajang

Figure 1:

Upper Langat River Basin Subbasins (Spatially Analysed Land Use Map of 1997)

Some of the other climatological properties of the basin are presented in Table 1. Table 1: Langat River Basin Climatological Properties (Average Amounts) Climatological Property Annual Rainfall (mm) Rainy Days (Days) Temperature (°C) Humidity (%) Amount 1521 – 2883 190 32 80

3.

DATA SET

The basic data set, required to develop an input database to assess the changes in streamflow regime, were land use and hydrometric data. Therefore, two land use maps have been prepared using available Landsat TM satellite images for the years 1997, and 2006 derived from remotely sensed data acquired from the Ministry of Agriculture, Malaysia. The spatial analysis was performed using ArcView GIS (ArcMap version 9.2) at watershed levels to define the land use changes over time.

3

K. Abbasnezhadi

Also, hydrometric data, gauged at the Kajang station (station No. 2917401, Sungai Langat) at the catchment outlet were obtained from the Department of Irrigation and Drainage (DID). Flow data covers a period of 12 years from 1994 to 2005 which covers almost the same period of time for which land use maps have been obtained (1997, and 2006). The reason behind such data selection would be discussed later. Regarding the hydrometric data, it should also be mentioned that some years show a number of missing data which was taken into consideration, and will be discussed later on as well. 4. METHODOLOGY

The methodology adopted in this study is summarized in Figure 2.

Analysis Stages

1st Stage Land Use Data Analysis Land-use Changes for 1997-2006

2nd Stage Flow Data Analysis Streamflow Trend for 1994-2005

3rd Stage Joint Analysis

Remote Sensing

Preprocessing

Image Processing

Landuse data

Acquisition of Streamflow Data

Flow Duration Analysis

Analysis of Mean Monthly and Annual Flow Discharges

Evaluation of the Hydrological Changes due to the Impact of Land Use Changes.

Figure 2:

Applied Methodology

The study was carried out in three stages; first, a database was established and land use maps were analysed to detect land use dynamics. Evaluation of the hydrological impacts due to land use changes and modifications was done by integrating Remote Sensing, Geographical Information System (GIS) and statistical evaluation and analytical methods. For more details about this stage see Appendix B. As a result, in the second stage, analytical assessment of streamflow data was accomplished; taking into account several statistical aspects of historic discharge data, including mean, maximum, and minimum annual and monthly discharge. Finally, a joint analysis of both stages was performed to evaluate the total hydrological changes which have occurred throughout the catchment, and a conclusion was produced on the impact of land use change on the streamflow discharge. 5. RESULTS AND DISCUSSION

As the first stage of the adopted methodology proposed, a temporal analysis was carried out on land use maps to arrive at the amounts of changes different land use types have gone through during the 1997-2006 period. 5.1 LAND USE ANALYSIS

As shown in Figure 3, two land use maps have been analysed incorporating ArcMap and the available satellite images.

4

Parallel Interaction of the Hydrologic Regime and the Spatio-Temporal Land Use Patterns

101° 50’ 30’’ E

101° 50’ 30’’ E

3° 11’ 0’’ N

3° 5’ 0’’ N

101° 50' 30” E

101° 50’ 30’’ E

Figure 3:

Land Use Maps of the Upper Langat River Catchment, 1997 (left), and 2006.

After some harmonisation, 7 different types of land use have been distinguished in these maps, as shown in Table 2. 5.1.1 1997-2006 Period

In this period, the maximum change was related to a sharp decreasing trend of 8% in the extent of vegetation areas which showed the disappearance of natural vegetation covers and forests (deforestation). Table 2: Different Land Use Types and the Corresponding Quotas during 1997-2006 Area (%) Land Use Type 1997 Vegetation Areas Forest Lands Wetlands Water Bodies Barren Lands Urban and Built-up Areas Open Spaces Total 21.53 56.41 0.25 1.29 1.20 18.07 1.25 100 2006 13.57 48.88 0.15 1.29 0.88 33.59 1.64 100 Total Change - 8.00 - 7.50 - 0.10 0 - 0.32 15.52 0.40 0 Net Change - 37.2 - 13.3 - 40.0 0 - 26.7 85.9 32.0 –

3° 5’ 0’’ N

1997

2006

3° 11’ 0’’ N

5

K. Abbasnezhadi

There are several factors accounting for this, such as changing of land use from forests to agricultural lands or into urban areas (urbanization), or it can be as a result of excessive logging of timber stocks. An increasing trend of 15.5% throughout the catchment (86% increase in the total area of urban and built-up areas) can be seen in the amount of land under urban and built-up areas. Transformation of agricultural areas into industrial and built-up areas can be a well defining factor behind such as increase in total area of urban and built-up land use type. Also the extent to which marshlands and wetlands have decreased by almost 40% can also be an alarming matter. In short, the development of two major agricultural crops, oil palm and rubber, is one of the primary causes of primary tropical forests loss. During the period under study, oil palm total area increased whereas rubber areas reduced. The development of oil palm has resulted in much loss of forestlands. On the other hand, some plantation areas have been changed into urban and built-up areas. 5.2 FLOW ANALYSIS

As the second stage of the adopted methodology proposed, to have a clear understanding of the overall changes of flow regime, the flow data have been divided into two time periods, which also include a time span well before the first land use map is acquired (1997). These two time spans are 1994-1997, and 2000-2005. The reason behind this categorization is to study the changes streamflow had experienced before and after land use changes have been studied. As mentioned, some missing data have been presented among the flow data obtained from the Kajang hydrometric station. For a number of years, the missing data can be neglected with some negligible error on the results, but for the 1998 and 1999 flow data the number of missing data constitutes almost all of the days for which data should have been available. As a result, in order to analyse different aspects of flow regime to arrive at a rational conclusion, these two years were not included in this study. 5.2.1 FLOW DURATION ANALYSIS

A flow duration curve describes the basin potential to afford different flow magnitudes. The relative amount of time flow requires for passing a site which is probable to equal or surpass a given rate of interest is extremely valuable for the designing purposes of a structures on a stream. For instance, a structure can be designed to perform well within some range of flows, such as flows that happen between 20% and 80% of all times (Oregon State University, 2002). Daily flow duration data for the two time periods that were considered for the Upper Langat River have been compared in order to determine the flooding and drying trends for the entire river catchment. Figure 4 illustrates the flow duration curves for the 1994-1997 and 2000-2005 time periods.
60
2000-2005 1994-1997

Q (m3sec-1)

40
Extends to maximum average daily flow of 212 m3sec-1 at 0.05%

20

0
0 20 40 60 80 100 % of Times Discharge Equalled or Exceeded

Figure 4:

Flow Duration Curve of Upper Langat River

The curves show an increase in the larger discharges for later time period (2000-2005), compared to the former one (1994-1997). The upward shift of the flow duration curve reveals a general surge in discharges for the latter period. Such observed variations in the flow duration curve and discharge rates suggest potential hydrologically related changes in the watershed, which can be due to land use changes; changes in rainfall intensity; changes in seasonal distribution of rainfall; and changes in rainfall duration. For a better view, the curves were plotted in logarithmic scale as shown in Figure 5.

6

Parallel Interaction of the Hydrologic Regime and the Spatio-Temporal Land Use Patterns

100
Log. (2000-2005) Log. (1994-1997)

Q (m3sec-1)

10

y = -7.90ln(x) + 36.41 R² = 0.97 y = -4.47ln(x) + 21.67 R² = 0.95

1
0 20 40 60 80 100

% of Times Discharge Equalled or Exceeded Figure 5: Flow Duration Curves (Log Intervals) For the Upper Langat River (1994-2005)

It can be realized that for discharges more than 3.55m3sec-1, the probability of flow exceeded than or equalled is higher for the flow data of 2000-2005. This, in turn, indicates the higher probability of flood occurrence for this time period, compared to the former one. As another example, while the flow duration curve for 1994-1997 period shows that the probability of flow being equalled to or exceeds 10m3sec-1 is almost 11%, the curve for 2000-2005 period suggests that the corresponding value has increased to 27% for the latter time period. The shape of a flow-duration curve in its upper and lower regions is particularly significant in evaluating the stream and basin characteristics. The shape of the curve in the high-flow region indicates the type of flood regime the basin is likely to have, whereas, the shape of the low-flow region characterizes the ability of the basin to sustain low flows during dry seasons (Oregon State University, 2002). Therefore, a very steep curve (high flows for short periods) would be normal for rain-caused floods on small watersheds which is more exposed for the 2000-2005 curve (Figure 4). On the other hand, in the low-flow region, the curve for 1994-1997 period – when compared with the one for 2000-2005 period – shows a more leveled curve, demonstrating that moderate flows are sustained throughout the former time period. 5.2.2 FLOW REGIMES

It is essential to differentiate between the effects on water yield (total streamflow) and flow regimes (the seasonal distribution pattern of streamflow) when the effects of land use changes are needed to be sorted out. The collective effects of both are given by the annual rates of streamflow discharge, and thus the real variations that have happened in the flow pattern during the relatively wet and dry periods might be concealed. To discern such seasonal changes within the annual cycle, changes in flow regimes should have first been investigated. Figure 6 shows the variations in flow regimes, based on mean monthly discharge during both periods. Each shows a different trend in annual total discharge. Also mean monthly rainfall given in Figure 6 shows that the distribution of rainfall is bimodal peaking in April and November each year. The Malaysian Peninsula is affected by two major monsoons; the Southwest Monsoon from late May to September, and the Northeast Monsoon from November to March. While the Northeast Monsoon brings heavy rainfall, particularly to the east coast states, the Southwest Monsoon which affects the Upper Langat River catchment normally signifies relatively drier weather. This fact justifies the behavior of mean monthly discharges from May until September which is relatively low-flow months. In the context of streamflow discharge, as shown in Figure 8, it can be observed that during the most prominent portions of low-flow period between May and September, and high-flow period between October and April, streamflow shows two unlike patterns for the two time periods. During low-flow period, it has not changed relatively during 2000-2005 time period, compared to 1994-1997 one. This is particularly illustrated by the 3 months moving average of rainfall and discharge. However, during high-flow period, a different pattern can be observed. Compared to 1994-1997, it has increased significantly during 2000-2005 period. This can be discussed in one way by strategies, resulting in land use changes with negative effects on streamflow discharge. Another point which can be concluded from Figure 6 is that the storage of rainfall on land has been changed, and this suggests a reduced watershed capacity to hold rainfall water rather than discharging it from 1994-1997 to 2000-2005.

7

K. Abbasnezhadi

25

Poly. (1994-1997) Poly. (2000-2005) Poly. (Mean Precip.)

350 300 250

20

15

P (mm)

200 R² = 0.90 150 100

10

5

R² = 0.70 50

0

0

Figure 6:

Mean Monthly Precipitation and Discharge during 1994-2005

It can be seen that during 1994-1997, the pattern of stream discharge throughout the year was rather more uniform when compared with that of 2000-2005 time period. Besides, Figure 8 obviously shows that during the low-flow period not much difference has happened between the flow behaviors of both time frames, but then during the high-flow period, streamflow discharges have become more intense for the latter time period which in turn can be justified in one way by land use change commanding effects on the stream behavior. To realise the primary grounds of such changes in the flow regimes, annual flow discharges have been analysed separately and subsequently during the following two time spans; i. ii. May – September (Low-Flow Period) October – April (High-Flow Period)

As shown in Figure 7, consecutive months with a similar pattern of changes have been plotted, considering the mean values of discharge for such consecutive months. During the low-flow period from May until September, streamflow discharge annually decreases by about 0.06m3sec-1. Thus, it can be observed that the basin has become less capable of maintaining its flow discharge along the main stream during the low-flow period. On the other hand, as it can be seen in Figure 7, during the high-flow period from October until April, streamflow discharges annually increase by almost 0.51m3sec-1. Therefore, it can be observed that the basin has become more incapable of holding rainfall water rather than discharging it through its main stream during high-flow months.
18 May - Sep Oct - April

15

Q (m3sec-1)

12

9

6

3

0

Figure 7:

Mean Annual Discharge during May-Sept and Oct-April

Q (m3sec-1)

R² = 0.92

8

Parallel Interaction of the Hydrologic Regime and the Spatio-Temporal Land Use Patterns

5.3

FLOW DISCHARGE – LAND USE INTERACTION

In Figure 8, it is shown that 86% growth of impervious area (urban and built-up areas) together with 13% decline of forest lands have increased the mean annual stream discharge by 48% (derived from the graph) from 1997 until 2006.
14 70 60 50 40 30 20 10 0

Mean Discharge (m3sec-1)

12 10 8 6 4 2 0

Mean Annual Discharge Impervious Area (%) Forest Land (%)

Figure 8:

Land Use versus Mean Annual Discharge Change 6. CONCLUSIONS

The applied GIS based methodology for the assessment of the effects of land use change on the streamflow regime of the Upper Langat River catchment in Selangor state of Malaysia has revealed that during high-flow months (October to April) in the 2000-2005 time period, most of the runoff is discharged from the catchment by flooding. The reason is laid upon the observed changes in forest as well as impervious land covers. However, it is clear that in order to produce an exact opinion on the effects of land use changes on the hydrologic regime of the basin in this time period, the effects of other prominent parameters such as mean annual rainfall should be regarded as an accounting factor as well. ACKNOWLEDGMENTS I would like to express my appreciation to Assoc. Prof. Abul Halim bin Ghazali for his support on contributions into this paper. Also, I would like to express my gratitude to anyone who helped me accomplish this project, my parents, my wife, and everybody else. REFERENCES Erle, E., and Pontius, R. (2007). Land use and Land-cover Change. Encyclopedia of Earth. Eds. Cutler J. Cleveland .Washington, D.C.: Environmental Information Coalition, National Council for Science and the Environment. http://www.eoearth.org/article/Land-use_and_land-cover_change Hadi, S., (2000). Malaysian Urbanisation and the Environment. Sustainable Urbanisation in the new Millennium. Environmental Management Programme. Bangi. pp. 76. Jahi, J., and Hasan, N. (1996). Federal-State Relationships in Environmental Management in Malaysia. Workshop on “Federal-State Relationships in Environmental Management”. ISIS Malaysia and MOSTE. Maksimovic, C., and Tucci, C. (2001). Urban Drainage in Specific Climate. IHP-V. Technical documents in Hydrology (Eds). No. 40, Vol.1. UNESCO, Paris.

Land-use Area (%)

9

K. Abbasnezhadi

Noorazuan, M.H., Rainis, R., Juahir, H., Sharifuddin; Zain, M., and Nazari, J. (2003). GIS Application in Evaluating Land use-Land Cover Change and its Impact on Hydrological regime in Langat River Catchment, Malaysia. http://www.gisdevelopment.net/application/environment/water/ma03039pf.htm Oregon State University, Civil, Construction, and Environmental Engineering Department. (2002). An Interactive Guide and Tutorial on Streamflow. http://water.oregonstate.edu/streamflow

10

Parallel Interaction of the Hydrologic Regime and the Spatio-Temporal Land Use Patterns APPENDIX A: Extra figures

Mean Monthly Discharge (As of the Mean Annual Flow)
250 200 150 2.5 2.0 1.5 1.0 0.5 0.0

%
100 50 0

Mean Monthly Discharge (1994-2005)
60 Mean MAX MIN 40 MEAN+ STDEV MEAN-STDEV

Q (m3sec-1)

20

0

Mean Annual Discharge
16 12 Q (m3sec-1) 8 4 0

σ = 7.34 m3sec-1

Ratio

11

K. Abbasnezhadi

Variation of Mean Annual Discharge (1994-2005)
15 12 Q (m3sec-1) 9 6 3 0

APPENDIX B: Spatio-Temporal Analysis Stage

Spatio-Temporal Analysis

1. Remote Sensing

2. Preprocessing

3. Image Processing

3. Land use Information

1. Restoration

2. Enhancement

3. Classification

Geometric Correction Haze Correction 1
3 4

FCC 3

Supervised

Unsupervised

PCA 4
FCC: False Color Composition PCA: Principal Component Analysis

NDVI 2

1 Brc = Bx − �Bxmin − B7min � Brc : Bx : Bxmin : B7min :

Haze corrected image Value of each band The minimum value of corresponding image The minimum value of band 7 of the image

2 NDVI =

NDVI: Normalized Difference Vegetation Index TM4 : TM spectral band 4 TM3 : TM spectral band 3

TM4−TM3 TM4+TM3

Sign up to vote on this title
UsefulNot useful