Land-Water Linkages in Rural Watersheds

Electronic Workshop
18 September – 27 October 2000

Background
Paper No. 1

Land use impacts on water resources

Ian Calder
Centre for Land Use and Water Resources Research
University of Newcastle upon Tyne, UK
E-mail: i.r.calder@newcastle.ac.uk

FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS

Rome, Italy
 Abstract

This paper aims to provoke discussion on land use and water resource impacts, particularly in
relation to rural watersheds in developing countries, with a view towards surfacing key issues,
identifying research needs and, ultimately, towards developing guidelines on instruments to
distribute costs and benefits arising from land use impacts on water resources amongst
upstream and downstream stakeholders in a watershed.

The paper is in three parts, which address and question:

1. The adequacy of our scientific knowledge in relation to the environmental processes

(biophysical/climatic) that determine land use impacts on water resources, summarizing
key points from a CGIAR SWIM Paper (Calder, 1998) and an Earthscan publication
(Calder, 1999; http://www.cluwrr.ncl.ac.uk).

2. The (often poor) connection between scientific knowledge and policy; the adequacy of the
decision-making and policy-making processes of national and international organizations
in relation to land use and water resources management; the self sustaining nature of
pseudo science myths in relation to land use and water resources and the interdependence
and interrelationships of stakeholders in relation to land use and water resources.

3. The adequacy of current management approaches and the need for consistent policies
towards land use and water resources management, development and poverty alleviation,
which are applicable from the local to the global scale.

1 Introduction

[IC-1] It is argued that to deal satisfactorily with land use impacts on water resources requires
an adequate scientific knowledge base, it needs this knowledge base to be connected
efficiently to the policy and decision making processes and it also requires the formulation of
land use and water management policies which are not only upwardly and downwardly
compatible, from the local rural watershed scale to the global scale, but are also consistent
with, and are developed alongside, other global and local policies relating to sustainability,
climate change, biodiversity, trade, food production and poverty alleviation.

[IC-2] This paper seeks to open the debate on where effort is still required to assist with the
management of land use and water impacts in rural catchments.

2 The Knowledge base – myth or reality

[IC-3] It has been argued (Calder, 1998, 1999) that the knowledge base that is being used for
land use and water resource decision making is often still based more on perceived wisdom
(myth) than science established reality. It is suggested that it is particularly necessary to purge

1
the myths relating to forests and water because, arguably, not only do forestation and
deforestation activities account for the largest area of land use change occurring on the planet
today, but they also account for some of the largest hydrological impacts. Seven “mother
statements” relating to forestry and agroforestry issues are considered, as a means to highlight
issues and identify gaps in our knowledge:

1. Forests increase rainfall?

2. Forests increase runoff?
3. Forests regulate flows?
4. Forests reduce erosion?
5. Forests reduce floods?
6. Forests “sterilize” water supplies - improve water quality?
7. Agroforestry systems increase productivity?

[IC-4] Two additional, in some cases related, issues are also raised in this paper as areas
where our understanding of the underlying processes and outcomes are uncertain:

8. Salinity control and downstream impacts

9. Land use and water related natural disasters – anthropogenic impacts and public
perceptions

[IC-5] The available literature and field experience related to these issues is considerable.
The titles referred to in the following sections are just examples and by no means exhaustive.
A fuller account of the seven “mother statements” is given in earlier publications (Calder,
1998, 1999) and only a summary is provided here together with relevant points made during
an informal internet debate, stimulated by the CGIAR’s Polex newsletter of December 1998
(contact David Kaimowitz d.kaimowitz@cgiar.org), making reference to papers by Calder
(1998) and Chomitz and Kumari (1998).

2.1 Forests increase rainfall?

[IC-6] Pereira (1989) states in relation to forests and rainfall:

"The worldwide evidence that high hills and mountains usually have more rainfall
and more natural forests than do the adjacent lowlands has, historically led to
confusion of cause and effect. Although the physical explanations have been
known for more than 50 years, the idea that forests cause or attract rainfall has
persisted. The myth was created more than a century ago by foresters in defence
of their trees… . The myth was written into the textbooks and became an article of
faith for early generations of foresters.

[IC-7] The overwhelming hydrological evidence supports Pereira’s view that forests are not
generators of rainfall. Yet this “myth,” like many others in forest hydrology, may contain a
modicum of truth that prevents it from being totally “laid to rest.” In earlier papers (Calder,
1998; 1999) it was argued that there is some evidence for land use controls on precipitation,
but often the magnitude of these effects are considerably less than is commonly imagined.
Giambelluca et al. (1999) also make the important observation that many of the recent Global
Circulation Model (GCM) predictions of reduced precipitation following forest clearance are

2
likely to be overestimates because replacement-vegetation ground surface parameters relating
to grass, rather than scrub or secondary re-growth forest (the more usual vegetation cover
following forest clearance), have been used in the simulation. Giambelluca et al. (1999)
claim, in relation to forests in Northern Thailand,

It is likely, therefore, that simulated reductions in precipitation in the region due

to deforestation will not be seen in model runs using more realistic scenarios of
post-deforestation land cover characteristics.

[IC-8] Perhaps the issue is best summed up by Bands et al. (1987), quoting from experience
in South Africa:

Forests are associated with high rainfall, cool slopes or moist areas. There is some
evidence that, on a continental scale, forests may form part of a hydrological
feedback loop with evaporation contributing to further rainfall. On the Southern
African subcontinent, the moisture content of air masses is dominated by marine
sources, and forestation will have negligible influence on rainfall and
macroclimates. The distribution of forests is a consequence of climate and soil
conditions – not the reverse.

[IC-9] Conclusion: Although the effects of forests on rainfall are likely to be relatively
small, they cannot be totally dismissed from a water resources perspective.

[IC-10] Research requirement : Further research is required to determine the magnitude of

the effect, particularly at the regional scale.

2.2 Forests increase runoff?

[IC-11] Based on process studies and from catchment experiments carried out throughout the
world, a new understanding has recently been gained of the evaporative regime in forests and
the way forests influence, and generally decrease runoff (contrary to widely accepted
folklore), as compared with shorter crops.

[IC-12] The difference between forest and short-crop evaporation is most marked in the
extremes of both very wet and very dry climates. In wet conditions, interception losses will be
higher from forests than from shorter crops primarily because of increased atmospheric
transport of water vapour from their aerodynamically rough surfaces. In dry (drought)
conditions transpiration from forests is likely to be greater because of the generally increased
rooting depth of trees compared with shorter crops, and trees' consequent greater access to
soil water.

[IC-13] The few exceptions lending some support to the folklore are:

1. Cloud forests, where cloud-water deposition may exceed interception losses.

2. Very old forests. Langford (1976) showed that following a bushfire in very old (200
years) mountain ash (Eucalyptus regnans) forest covering 48% of the Maroondah
catchment, one of the water supply catchments for Melbourne, Australia, runoff was

3
 reduced by 24%. The reason for this reduction in flow has been attributed to the increased
evaporation from the vigorous re-growth forest that had a much higher leaf area index
than the former very old ash forest.

[IC-14] Conclusion: Notwithstanding the exceptions outlined above, catchment experiments

generally indicate reduced runoff from forested areas compared with areas under shorter
vegetation (Bosch and Hewlett, 1982; Hamilton, 1987).

[IC-15] Caveat: Information on the evaporative characteristics of different tree species+soil

type combinations are still required if evaporation estimates with a certainty of over 30% are
required. In both temperate and tropical climates, evaporative differences between species and
soil types are expected to vary by about this amount.

2.3 Forests regulate flows – increase dry season flows?

[IC-16] Although it is possible, with only a few exceptions, to draw general conclusions with
respect to the impacts of forests on annual flow, the same cannot be claimed for the impacts
of forests on the seasonal flow regime. Different, site-specific, often competing, processes
may be operating and the direction, let alone the magnitude of the impact, may be difficult to
predict for a particular site.

[IC-17] From theoretical considerations it would be expected that:

1. Increased transpiration during dry periods will increase soil moisture deficits and reduce
dry season flows.

2. Increased infiltration under (natural) forest will lead to higher soil water recharge and
increased dry season flows.

3. For cloud forests, increased cloud-water deposition may augment dry season flows.

[IC-18] There are observations (Robinson, Moore and Blackie, 1997) which indicate that for
the uplands of the UK, drainage activities associated with plantation forestry increase dry
season flows both through the initial de-watering and in the longer term through alterations to
the hydraulics of the drainage system. The importance of mechanical cracking associated with
field drainage and its effects on drainage flows has been highlighted by Robinson, Ryder and
Ward (1985), whilst the work of Reid and Parkinson (1984) indicates that landform and soil
type may sometimes be the dominant factors determining soil moisture and drainage flow
response.

[IC-19] Bruijnzeel (1990) discusses the impacts of tropical forests on dry season flows and
concludes that the infiltration properties of the forest are critical for how the available water is
partitioned between runoff and recharge (leading to increased dry season flows).

[IC-20] There are a number of observations from South Africa that indicate that increased dry
period transpiration following forestation with pine or eucalyptus species will significantly
reduce low flows (Bosch, 1979; van Lill, Kruger and van Wyk, 1980; Scott and Smith, 1997).
Scott and Lesch (1997) also report that on the Mokobulaan research catchments under

4
Eucalyptus grandis, the stream flow completely dried up by the ninth year after planting.
When the eucalypts were clearfelled at age 16 years, perennial stream flow did not return for a
further five years. They attributed this large lag time as being due to very deep soil moisture
deficits generated by the eucalypts, which required many years of rainfall before field
capacity conditions could be re-established and recharge of the groundwater aquifer and
perennial flows could take place.

[IC-21] Conclusions : Forestation will not necessarily increase dry season flows. Competing
processes may result in either increased or reduced dry season flows. Effects on dry season
flows are likely to be very site specific.

[IC-22] Caveat: The complexity of the competing processes affecting dry season flows
indicates that detailed, site-specific models will be required to predict impacts. In general, the
role of vegetation in determining the infiltration properties of soils remains poorly
understood, as it affects the hydrological functioning of catchments through surface runoff
generation, recharge and high and low flows and catchment condition. Modelling approaches
that are able to take into account vegetation and soil physical properties, including the
hydraulic conductivity and water content properties of the soil, and possibly the spatial
distribution of these properties, will be required to predict these site-specific impacts.

2.4 Forests reduce erosion?

[IC-23] As with impacts on seasonal flows, the impacts of forests on erosion are likely to be
site-specific, and again, many, and often competing processes, are likely to be operating.

[IC-24] In relation to beneficial impacts, conventional theory and observations indicate that:

1. The high infiltration rate in natural, mixed, forests reduces the incidence of surface runoff
and reduces erosion transport.

2. The reduced soil water pressure and the binding effect of tree roots enhance slope
stability, which tends to reduce erosion.

3. On steep slopes, forestry or agroforestry may be the preferred option where conventional
soil conservation techniques and bunding are not appropriate.

[IC-25] Adverse effects, often related to forest management activities, may result from:

1. Bad logging techniques, which compact the soil and increase surface flow.

2. Pre-planting drainage activities, which may initiate gully formation.

3. Wind throw of trees and the weight of the tree crop reduce slope stability, which tends to
increase erosion.

4. Road construction and road traffic, which can initiate landslips, gully formation and the
mobilization of sediments.

5
5. Excessive grazing by farm animals, which leads to soil compaction, the removal of
understorey plants and greater erosion risk.

6. Splash induced erosion from drops falling from the leaves of tree canopies, particularly if
the litter layer and understorey are missing.

[IC-26] The effects of catchment deforestation on erosion, and the benefits gained by re-
foresting eroded catchments will be very dependent on the situation and the management
methods employed. Quoting Bruijnzeel (1990):

In situations of high natural sediment yield as a result of steep terrain, high

rainfall rates and geological factors, little, if any influence will be exerted by man.

[IC-27] In the Himalayas, for example, there is evidence that a large proportion of suspended
sediments in the rivers is contributed by big, deep, geologically induced landslides, which
occur on any type of land cover, including forests (Galay, 1985). Also, in situations where
overland flow is negligible, in drier land, little advantage will be gained from forestation.
Versfeld (1981) has shown that at Jonkershoek, in the Western Cape of South Africa, land
cover has very little effect on the generation of overland flow and soil erosion.

[IC-28] In more intermediate conditions of relatively low natural rates of erosion and under
more stable geological conditions, man-induced effects may be considerable. In these
situations catchment erosion may well be hastened by deforestation and there may also be
opportunities for lessening erosion by well-managed forestation programmes. Even in these
situations, forestation should not necessarily be seen as a quick panacea. In heavily eroded
catchments, such as those on the slopes of the Himalayas, so much eroded material will have
already been mobilized that, even if all the human-induced erosion could be stopped
immediately, it would be many decades before there would be any reduction in the sediment
load of the rivers (Pearce, 1986; Hamilton, 1987).

[IC-29] The choice of tree species will also be important in any programme designed to
reduce erosion. Recent theoretical developments and observations (Hall and Calder, 1993;
Calder, 1999) confirm that drop size modification by the vegetation canopies of trees can be a
major factor leading to enhanced splash-induced erosion. These observations indicate that tree
species with larger leaves generally generating the largest drop sizes. The use of large-leafed
tree species, such as teak (Tectona grandis) in erosion control programmes would therefore be
ill advised, especially if there is any danger of understorey removal.

[IC-30] Conclusions : It would be expected that competing processes might result in either
increased or reduced erosion from disturbed forests and forest plantations. The effect is likely
to be both site- and species-specific. For certain species, e.g. Tectona grandis, forest
plantations may cause severe erosion. It is a common fallacy that plantation forests can
necessarily achieve the same erosion benefits as natural forests. Forest cover as such does not
guarantee low rates of erosion – the forest quality (density of trees, quality of the lower
canopy layers, availability of surface litter, etc.) is an equally – if not more – important factor
(Hamilton, 1987). Smyle (2000, personal communication) has suggested that the erosion rates
in undisturbed natural forest might be considered to represent a “natural baseline” or
“background” erosion rate against which the erosion rates from all other land uses could be

6
compared. The use of such an index may well be of value in land use management and the
design of realistic erosion control programmes.

[IC-31] Caveat: Although conventional erosion modelling methods, such as the Universal
Soil Loss Equation (USLE) (USDA, 1961), provide a practical solution to many problems
associated with soil loss from agricultural lands, it may not be adequate for the prediction of
erosion resulting from forestation activities. Process understanding of the erosive potential of
drops falling from different tree species is not adequately appreciated, and soil conservation
techniques related to vegetation type, soils and slope characteristics have not yet been fully
developed.

2.5 Forests reduce floods?

[IC-32] It is a widely held view, propagated by foresters and the media, that forests are of
great benefit in reducing floods. Disastrous floods in Bangladesh and northern India are
almost always associated with “deforestation of the Himalayas”; similarly in Europe, floods
are often attributed by the media to “deforestation in the Alps.” However, hydrological
studies carried out in many parts of the world – America (Hewlett and Helvey, 1970), South
Africa (Hewlett and Bosch, 1984), UK (Kirby, Newson and Gilman, 1991; Johnson, 1995)
New Zealand (Taylor and Pearce, 1982) and Asia (Bruijnzeel and Bremmer, 1989; Ives and
Messerli, 1989; Hofer, 1998a, 1998b) – show little linkage between land use and storm flow,
and therefore do not support this view. From theoretical considerations it would be expected
that interception of rainfall by forests reduces floods by removing a proportion of the storm
rainfall and by allowing the build up of soil moisture deficits. These effects would be
expected to be most significant for small storms and least significant for the largest storms.

7
 Box 1. Extracts from an internet debate: Forests and landslips

Ian Cherret (working on an FAO project, Honduras), discussing Hurricane Mitch:

Calder’s comment on management activities is very relevant [that management activities
associated with forestry, such as cultivation, drainage, road construction, and soil compaction
during logging, are more likely to influence flood/erosion response than the presence or absence of
the forests themselves]. This would be a good time to do field research in Honduras - Choluteca
Basin - where first impressions are that extensive landslides that did so much damage (a rough
estimate of level of sediment carried by flood waters running through Tegucigalpa at their height is
15-17%). The impression is also that landslides were concentrated where mature trees had been
cleared (only shallow roots holding the saturated soil) or where pine woods had suffered extensive
burning last May (impoverished soils). This deserves investigating. Recent research by Texas A&M
on an USAID-funded project in the worst-hit basin suggests that soil loss through landslides is at
least as important in Central America as loss through runoff and wind (Technical Bulletin No. 98-2,
1998). And the key to landslide control is deep roots provided by large trees.

Our experiences do not contradict the conclusions of Calder or Chomitz (relating to the importance
of management activities) but they do indicate that part of the need for further research requires
unravelling relationships such as that of SOIL and vegetational cover as opposed to TREES per
se.

Jim Smyle (Natural Resources Specialist, Project RUTA, Costa Rica):

The second topic (relating to lack of vegetation causing landslips) can be bit more thorny, as it can
be governed by very local factors, and those factors may have little to do with vegetation. For
example, the Choluteca basin had a tremendous amount of landsliding. It is also heavily
deforested. It also appears to have a geology and soil type which makes it highly susceptible to
land sliding. Deep-rooted trees could be the answer ... though in Hawaii's unconsolidated upland
soils, the weight of forest on hill slopes often makes them MORE susceptible to sliding.

Ian Cherret (working on an FAO project, Honduras), discussing Hurricane Mitch:

In the watershed Lempira Sur in Rio Mocal, serious landslides occurred in the upper watershed on
mount Celaque, the highest peak (2 900 m) in Honduras, at over 2 000 metres! They were
deforested areas though - a thousand people have had to be relocated because of just one
landslide. Clearly the underlying geology has something to do with propensity to landslides. I
believe that there is a need for a more scientific study looking at factors such as slope, cover,
geomorphology and rainfall pattern.

[IC-33] The high infiltration rates under natural forests also serve to reduce surface runoff
and flood response. Certain types of plantation forests may also serve to increase infiltration
rates through providing preferential flow pathways down both live and dead root channels.
(Through the use of border trees around agricultural fields subject to surface runoff generation
there may be some prospect for runoff and flood response mitigation whilst not introducing
excessive evaporative losses from wide expanses of trees.)

[IC-34] However, field studies generally indicate that it is often the management activities
associated with forestry – cultivation, drainage, road construction (Jones and Grant, 1996),
soil compaction during logging – that are more likely to influence flood response, rather than
the presence or absence of the forests themselves.

[IC-35] In the “Himalayan Dilemma” (Ives and Messerli, 1989) it is suggested that the scale-
approach is a key issue in the understanding of highland-lowland linkages. Whereas on a
micro-scale (small watershed) the effects of human interventions such as forest cutting can be
directly documented in terms of higher discharge peaks or higher sediment load, on a large-
scale (e.g. Ganga-Brahmaputra-Meghna Basin in Bangladesh) natural processes are dominant,

8
and the impacts of human activities in the Himalayas are neither detectable nor measurable.
This does not relieve the mountain inhabitants of their responsibility to use the environment in
a sustainable manner, however. Lauterburg (1993) states:

Forestation of mountain watersheds and extensive soil conservation measures are

valuable for the sake of the hill farmers, if appropriately carried out. It is
potentially disastrous, however, for foreign aid agencies and national government
authorities to undertake such activities with the conviction that they will solve
problems on the plains.

[IC-36] Conclusions: For the largest, most damaging flood events there is little scientific
evidence to support anecdotal reports of deforestation as being the cause.

[IC-37] Caveat: Carefully conducted controlled catchment experiments with various climate-
soil-species combinations will be required to resolve this issue, but species impacts are
probably not as significant as often portrayed. Management activities are more likely to be
paramount. At smaller spatial scales, there is a greater likelihood of land use affecting both
peak storm flow and the time-to-peak.

Box 2. Extract from an internet debate: Deforestation-sedimentation-flooding

Bruce Aylward (Environmental economist):

One complicating factor in this regard that is not often considered in flooding is the potential role of
sedimentation of waterways. I am not saying it is the case, but what if the river bottom is now X
inches/feet higher than it used to be due to the accumulation of sediment and bed-load - one source
of which would be forest conversion/logging, etc? That said, I expect the damages have much more
to do with population/development trends and too much rain. Question is, does climate change play
into the latter ... in which case forests do play a role.
I have been waiting for someone to blame the disasters associated with Hurricane Mitch on
deforestation as well, and was partially rewarded by a commentary sent into the Oregonian:

Contrary to the television images, the devastation wrought by Mitch was no mere act of
God but a far more human tragedy. Misguided government policies and poor farming
practices - the two are interrelated - had already pushed the region to the brink of
ecological collapse. The torrential rains only gave it a final push …

The article then goes on to repeat (false) conventional wisdom on forests and soil/water
conservation.

Ian Cherret (working on an FAO project, Honduras), discussing Hurricane Mitch:

[Indicating levels of sediment in flood waters following landslips in Choluteca Basi] a rough estimate
of level of sediment carried by flood waters running through Tegucigalpa at their height is 15-17%.

Jim Smyle (Natural Resource Specialist, Project RUTA, Costa Rica):

Yes, the flooding impacts were probably made worse by channel aggradation (i.e., river bottoms
raised by the sedimentation caused by deforestation and upland land use changes) in some reaches
of rivers, but I would bet this was less of a problem than human encroachment in flood plains...

9
 Box 3. Extract from an Internet debate: Forests and Floods – The Scale Issue

From Robert K. Simons (Hydrological consultant):

On this subject, our experience has been that deforestation causes increased water and sediment
runoff. We conducted detailed data collection studies and developed computer models of
watershed water and sediment runoff specifically related to timber activities for the USDA Forest
Service on experimental watersheds. These were relatively small in size. Given the fact that there
is increased water and sediment runoff for small watersheds, it is difficult to believe that there is no
increase for larger watersheds.

I suspect it is more difficult to determine for larger watersheds because runoff from large
watersheds integrate the cumulative effect of the entire watershed, of which only a relatively small
percentage may have been affected by deforestation. In addition, hydraulic attenuation (which
tends to result in decreased flood peaks but longer timed base hydrographs) may mask the
increased flow. Based on the type of hydrologic modelling that could be conducted, one could
develop some sense of the magnitude of increased runoff effects on larger watersheds.

From Ken Chomitz (Development Research Group, World Bank):

We really need to run these models for large basins using detailed info about the geographical
size, intensity, and movement of large storms. There is evidence that deforestation can cause
flooding in small basins.

Bruijnzeel's argument is that small storms average out over large basins; and that storms large
enough to cover an entire basin at once tend to be so intense as to saturate the ground, so that
runoff is the same with and without forest. The case is by no means closed and any info would be
useful.

From Ian Calder (CLUWRR):

... deforestation generally leads to increased annual and often seasonal runoff, the deforestation
process leads to increased erosion but the effect of deforestation on flood peaks, especially the
largest flood peaks, is probably marginal (and more related to the deforestation process and other
forest management activities than the presence or absence of the forest itself). I would imagine
that the USDA model would also show that as the size of the flood event increases the effects of
land use would become less important. I am still unaware of any hydrological (as opposed to
anecdotal) evidence that establishes a link between major floods and deforestation.

2.6 Forests “sterilize” water supplies – improve quality?

[IC-38] Forests were historically the preferred land use for water supply catchments because
of their perceived “sterile” qualities associated with an absence of livestock and an absence of
human activities. More recently, the generally reduced fertilizer and pesticide applications to
managed forests and forest plantations compared with agricultural lands has been regarded as
a benefit with regard to water quality of runoff and recharge. Reduced soil erosion from
undisturbed or well-managed natural forests can also be regarded as a benefit.

[IC-39] Offsetting these benefits, management activities (cultivation, drainage, road

construction, road use, felling) are all likely to increase erosion and nutrient leaching.
Furthermore, deposition of most atmospheric pollutants in forests is higher because of the
reduced aerodynamic resistance of forest canopies compared with those of shorter crops. In
high pollution (industrial) climates, this is likely to lead to both long-term acidification of the
catchment and acidification of runoff.

10
[IC-40] Conclusions : Except in high pollution climates, water quality is likely to be better
from forested catchments. Adverse effects of forests on water quality are more likely to be
related to bad management practices rather than the presence of the forests themselves.

[IC-41] Caveat: Studies may still be required to determine the magnitude of the impacts for
specific sites and the means to minimize adverse impacts.

2.7 Agroforestry systems increase productivity?

[IC-42] Agriculturists have long recognized the productivity benefits that can be obtained
from crops by avoiding resource constraints to growth, particularly in relation to water and
nutrients, and, in some situations, also with regard to light and carbon dioxide. Increasingly,
the tradeoffs between increased production from irrigation schemes are being assessed with
respect to the costs to downstream users of often both reduced and poorer quality water
resources. Perhaps less well appreciated is that systems that increase productivity through
crop mixtures and agroforestry are likely, if successful, also to have downstream impacts
through increased resource (water resource) use. Arguably, there is a case for cost:benefit
analysis that takes into account upstream-downstream benefits not only for forestry and
irrigation systems but also for agroforestry systems, when these systems significantly affect
downstream water resources.

[IC-43] Farmers are well aware of the productivity benefits that can sometimes be gained by
mixing different agricultural crops. When a crop such as pigeon pea is mixed with sorghum or
maize, much higher production can be obtained compared with pure stands these species on
the same area. When productivity of the mixture is superior to that of sole cropping, it is seen
that the mixture is overyielding and that complementarity has occurred. Requirements for
complementarity are that the crops, together in a mixture, make better use of resources (water,
light, nutrients and carbon dioxide) than when supplies of these resources are limited.

[IC-44] It is thought that the increased productivity that has been recorded from mixtures of
agricultural crops is largely the result of temporal complementarity, especially when one crop
is an annual and the other a biennial, or a result of, for example, nitrogen fixation or other
factor or condition created by one of the crops. Mixtures with pigeon pea, grown either as an
annual or biennial legume, exhibit both and has been found to show complementarity with
many species, such as maize, sorghum, groundnut and cowpea (Ranganathan and de Wit,
1996).

[IC-45] Agroforestry is built on the belief that mixtures of agricultural crops and trees can
also lead to increased productivity, the belief that it is possible to find overyielding mixtures
of crops and trees which, when combined, would have a higher yield than when grown
separately on the same land area.

[IC-46] However it is now becoming evident that the claims in support of the beliefs, and the
trials on which they are based, are flawed. Ong and colleagues, from detailed studies of
resource capture between both intercropping and agroforestry systems, carried out at the
International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) sites at
Hyderabad in India and the International Centre for Research in Agroforestry (ICRAF) sites

11
at Machakos in Kenya, have developed a scientific framework and process understanding
which is at variance with claims for increased biomass from agroforestry systems. These
experimental studies suggest that, with agroforestry systems, often the best that can be
achieved is near to neutral productivity difference.

[IC-47] Ong et al. (1991, 1996), by the application of rigorous scientific reasoning, have also
identified some of the pitfalls that have trapped agroforestry researchers into thinking that
complementarity and overyielding systems had been achieved. These have arisen mainly
because the control plots containing the sole crop stands have not been under optimal
conditions, so that the mixture receives a favourable bias:

1. For the convenience of statistical analysis, the same plant densities are used in both the
sole crops and the mixtures, although, to achieve optimal productivity in the sole stand,
higher densities would be required.

2. The mixture and sole crop are managed identically, even though this management may
result in sub-optimal productivity in the sole crop. For example, pruning, which has been
used in the mixture to reduce competition from the trees and to return nutrients to the
system, would not produce optimal productivity if applied to the control plot.

3. Plot sizes may be too small, allowing tree roots from the mixture or the sole tree plots to
penetrate into the plots of the sole agricultural crop and reduce their yields (van
Noordwijk et al., 1996). Ong (1996) has shown from studies at Machakos, Kenya, that the
roots of the tree Leucaena leucocephala can reduce the yield of maize 5 m away within
two years of growth.

[IC-48] One of the fundamental differences between agroforestry systems and intercropping
systems is that the tree component in an agroforestry system, after the initial establishment
period, has a well-developed and deep root system. Opportunities for spatial complementarity
of below-ground resources are therefore limited because the tree roots tend to exploit the
whole root zone.

[IC-49] Spatial complementarity of above-ground resources, particularly in relation to light

capture, is however achievable with tree-crop mixtures. Rao, Sharma and Ong (1990) have
shown that at the ICRISAT Hyderabad site a mixture of Leucaena leucocephala and millet
increased the light capture above that of a sole millet crop. Yet this improved light capture did
not result in increased biomass production of the mixture – the biomass produced by the trees
was essentially equal to the reduced biomass yield of the crop. This lack of improvement has
been explained (Cannell, Mobbs and Lawson, 1998) in terms of the photosynthetic processes
operating in trees (all C3 type) that are less efficient in their light-to-biomass conversion
efficiencies than crops, such as millet, which are of C4 type and will have much higher
efficiencies. Even though greater resource capture is achieved, this does not translate into
higher total productivity.

[IC-50] It is now becoming apparent that trees in agroforestry systems will generally lead to a
reduction in biomass of the associated crop (Ong, 1996) and neutral total biomass production
is usually the best that can be expected.

12
[IC-51] Modelling studies also support this view, vide Cannell, Mobbs and Lawson (1998),
who, through the use of a process-based agroforestry model that takes into account
competition for light and water (but not nutrients), were able to simulate the growth of a
sorghum and tree crop mixture under different climatic conditions. Their conclusions can be
summarized as:

1. at sites with less than 800 mm rainfall, maximum total site biomass production was
obtained with a sole crop, without overstorey trees;

2. at sites with 800-1000 mm rainfall, neutral biomass production was obtained with a
mixture;

3. at sites with >1000 mm rainfall, biomass production would be increased with a mixture,
provided the Leaf Area Index (LAI) of the trees was greater than 0.25, but this increase in
overall production would be at the expense of a 60% reduction in sorghum grain yield;
and

4. any decrease in crop yields due to tree competition will automatically increase the
frequency of years with poor yields and threaten food security.

[IC-52] Therefore, for low-rainfall sites, sole crops would clearly be the best option and, even
at higher-rainfall sites, to achieve higher total biomass production it would be necessary to
accept a huge – 60% – reduction in the sorghum crop yield. Cannell, Mobbs and Lawson,
(1998) make the point that “the biomass produced by the trees must be of considerable value,
relative to that of sorghum grain, for this sacrifice in yield to be worthwhile.”

[IC-53] The “holy grail” of agroforestry – a tree species that has roots at depth that can
exploit deep soil resources of water and nutrients but with few roots in the surface layers to
offer competition to shallow-rooted agricultural crops – is still being sought. But, even if it
ever were discovered, the spatial complementarity that would be achieved would not be
without costs. As for overyielding intercrops, the productivity gains would be at the expense
of increased resource use and, when the resource is water, any increased productivity gains
would need to be assessed in relation to the (marginal) cost of the extra water consumed.

[IC-54] The recent research findings discussed above have far-reaching implications for the
practice of agroforestry. They show, contrary to the “mother statements” underlying much of
agroforestry practice, that there are in fact few opportunities for gains in productivity by
mixing trees with agricultural crops.

[IC-55] Conclusions : Despite the claims made by overenthusiastic agroforesters, there is

little scientific evidence to show that enhanced productivity can be achieved in agroforestry
systems. Growth of the woody component will virtually always be associated with a decrease
in biomass and value of the associated annual agricultural crop. The huge investment in the
development and demonstration of agroforestry systems purporting to increase productivity
might not have been wasted had more attention been paid by development workers to the
indigenous knowledge of local people. When high yielding crop systems are achieved, it
should also be borne in mind that the productivity gains are being achieved at the expense of
increased resource use and, when the resource is water, any increased productivity gains

13
should be assessed in relation to the (marginal) cost of the extra water consumed, and the
impacts on downstream users of the water. These impacts need to be taken into account for all
high-water-use crop systems, whether forestry, agroforestry or irrigated agriculture.

[IC-56] Caveat: Although it would be expected that close proximity competition from the
woody component in agroforestry systems would generally prevent productivity benefits,
there might well be achievable synergies in agricultural crop and tree crop systems that rely
on rotations or tree crop fallows. With these systems, it is possible that the deep-rooted nature
of most trees may, although consuming deep soil-water reserves, bring to the surface through
leaf fall, nutrients that are located at depths greater than annual crops can access. A rotation
with an annual crop may then allow these nutrients to be utilized by the crop whilst the crop
itself, through having a less-developed root system, will not be able to utilize all the rainfall,
and some water will be available to recharge deeper layers in the soil that the tree crop had
previously depleted.

2.8 Salinity control in agricultural systems and downstream impacts

[IC-57] It is widely recognized that salinity control and the avoidance of waterlogging are
important considerations, not only in salinity-prone dryland farming but also in the design of
irrigation schemes. The failure of salinity control measures may lead not only to land being
rendered useless for productive purposes, but also lead to salinization of watercourses, with
consequent impacts on downstream users.

[IC-58] Deficiencies in our knowledge of the environmental processes determining

salinization are probably not a significant constraint in finding solutions to the problem. The
quotation by Robertson (1996) illustrates the frequently occurring situation worldwide, where,
when land use and water resources decisions need to be taken promptly and where there is no
lack of understanding of the processes, the issues are often fudged and deferred.

Salinity and its avoidance or management has been an enigma in Australia. On

one hand there has been a wide understanding of the inherently salty nature of the
environment and the inevitability of salinity. On the other, there has been a high
propensity to ignore the problem and believe that for some reason salinity will not
develop in any specific area.

[IC-59] Deferral of decision making in these circumstances leads to progressive salinization

of watercourses, destruction of the river ecology, and the lower reaches of river systems
becoming unfit for human use or irrigation. Three types of options have been recognized as
being beneficial in controlling salinity problems in dryland agriculture and irrigation schemes:
land management and engineering solutions, which are discussed here, and economic and
policy instruments, which are discussed in the last section of the paper.

2.8.1 Land management approaches to salinity control

[IC-60] In salinity-prone dryland areas, replanting tree and shrubs in the recharge areas of a
catchment will increase the evaporation, through increased interception and probably also

14
increased transpiration, and reduce recharge and groundwater levels. This will alleviate
salinity problems in seepage areas by both lowering watertables and lowering the level at
which seepage takes place, hence reducing the land area affected, and also by reducing the
volume of seepage waters. In areas where dryland salinity problems are not extreme, changes
in cropping pattern have been advocated. In Australia, research is now being conducted to
determine agricultural crop combinations that reduce recharge (Zhang et al., 1999) and, for
areas where salinity problems are more acute, combinations of agricultural and tree crops.

[IC-61] Land management has also been suggested as a method for salinity control in
irrigation schemes. In situations where water rather than land is the primary constraint, land
management involving “dry drainage” has been suggested (Gowing and Wyesure, 1992) as an
alternative to the traditional engineering approach. The serious downside to the dry drainage
approach, which involves perhaps up to one-half of the land area, rather than the watercourse,
becoming the sink for saline wastewaters, is that this “sacrificed” land becomes essentially
lost to productive use. On this sacrificed land surface, evaporation allows the build up and
concentration of salts in the soil profile. In conjunction with the dry drainage approach, the
use of shallow saline groundwater as a source of water for crop growth, rather than lowering
groundwater tables through pumping and then re-applying fresh water as irrigation water
following traditional engineering practice, has been advocated as having additional benefits.
The loss of land for productive use clearly does not make the dry drainage option particularly
attractive and cannot be regarded as a "sustainable" solution to land management. However,
perhaps in some circumstances, where salinization has already occurred, the approach may
have some benefits as compared with costly conventional engineering solutions.

2.8.2 Engineering solutions

[IC-62] Various engineering solutions have been proposed. These include the interception of
saline groundwaters and the diversion of the excess water (returns) from irrigation schemes to
evaporation pans or directly to rivers. Direct pumping of groundwater to lower groundwater
tables has also been advocated in both dryland agricultural areas and within irrigation
schemes, but a major problem with these solutions is the disposal of the saline effluents.
Disposal to rivers is the common option, but often serves only to salinize the vital water
supply for the less fortunate users who happen to be located downstream. If downstream users
also drain their field or pump groundwater in a similar fashion, rivers will undergo
progressive salinization until the ecology of the river is destroyed and lower reaches become
unfit for human use or irrigation. Piping the effluents for discharge into the sea has been
considered in some locations, but is generally thought to be not cost effective.

2.9 Land use and water related natural disasters – anthropogenic impacts and public
perceptions

[IC-63] Many development organizations have recently questioned perceived wisdom in

connection with the natural disasters of Hurricane Mitch and the floods in China. The type of
questions asked were:

15
Is there any evidence that the change in nature and/or frequency of natural disasters has been
caused by climate change or environmental degradation? Is there any evidence that their
impact has been greater because of environmental degradation, and if so, what types of
degradation?

[IC-64] Clearly these are important questions to address. Aid organizations are called on to
assist in the relief of natural disasters, and funds allocated for preventive measures could
conceivably be a more effective use. The media often cite anthropogenic impacts as the cause
of natural disasters, often reflecting the interests of one or other pressure groups whose
interests will be served by their “solutions.” The Financial Times article of 11/12 March
2000 by Victor Mallet argues in relation to the Mozambican flood of 2000 that “Drainage
schemes, overgrazing and the spread of concrete in Mozambique’s neighbours have destroyed
their land’s ability to absorb heavy rains.” The conversion of wetlands to agricultural lands is
also given as another reason for the flooding. At discussions at the World Water Forum in the
Hague on the Mozambican floods, the Dutch solution of dam building was evident: “But
most of all it’s very important to get the help of people with expertise who know how to build
dams to prevent future flooding." (see http://www.worldwaterforum.org)

[IC-65] A recent informal internet debate, stimulated by the CGIAR’s Polex newsletter of
December 1998, referring to papers by Calder (1998) and Chomitz and Kumari (1998),
involving representatives of the World Bank, CGIAR, international consultants and research
organizations, has highlighted not only the inadequacy in our scientific knowledge in relation
to the environmental processes and anthropogenic influences that may lead to and exacerbate
natural, water-related, disasters, but also another important factor: the public perception of
disasters. Perhaps the public perception of the increased frequency and severity of water
related disasters may be as much (if not more) related to increased media coverage, higher
world populations and higher numbers of people at risk (particularly in relation to the
increased numbers of people living in flood plains and subject to flood risk) as to climatic
change or land use change impacts? Understanding the role of indigenous knowledge may be
a key factor in comprehending the public perception of natural disasters (Barr, 1998).
Comments from this debate, which raised three new issues (or sub issues), are given below
with permission of the authors.

3 Connecting science to policy

[IC-66] For land use related research to have any value it has to be connected to the decision
making and policy making process. Traditionally, the culture of the research community,
which has been fostered and is still being fostered by research councils, is to publish research
findings in tightly defined disciplinary, peer-reviewed research journals - the number of
research papers published and the number of citations received determines the credit they
achieve. Many environmental research scientists still operate under the happy delusion that
once published, their findings will also shortly be taken up and made use of in the
development of environmental policy. Sadly, this is not often the case. The readers of
research journals are generally fellow researchers, not managers and decision makers.
Decades may elapse, particularly it seems in relation to land use and water resource issues,

16
before research findings trickle into policy. Arguably, addressing the reasons for this lack of
connection between science and policy might be the single most effective use of resources in
furthering land use and water resources management at all scales.

[IC-67] The problem is not just one of the developing world. The UK Government’s White
Paper on Rural England (DOE, 1995) called for a doubling of lowland forests, mainly for
amenity considerations, without demonstrating any awareness of the water resource
implications of such a large change in land use. Although the knowledge base for
understanding the reasons for dryland salinity in Australia have been known for over 50 years
(viz. the removal of forests reduces evaporation and saline water tables rise to produce saline
seeps) forests are still being cut in the dryland areas and, to aggravate matters further,
forestation is occurring in the headwater areas, thus reducing dilution flows. In Spain
forestation of the Pyrenees is being promoted by the Spanish Government to improve
downstream water resources. In Panama, Law 21, (with USAID support) has been passed,
which promotes forestation of the Panama catchments as a means of enhancing flows and
improving the functioning of the canal.

[IC-68] With respect to the need for water in order to increase the capacity of the Panama
Canal, it is estimated that if 1 000 ha/year of deforested land in the watershed was reforested,
it would not be necessary to construct an additional dam proposed for the Rio Ciri (CEASPA,
1997), based on calculations made by Nathan/Intercarib S.A., 1996, the consulting firm which
developed the basis for Law 21.

[IC-69] Where does the problem lie? With the researchers, for not making their results
available to a wider community of researchers operating in the different disciplines that all
relate to environmental research and decision making, and to the wider public and decision
making community? With the research councils, which still promote single disciplinary
research and where no “impact factor” is ascribed to any "real world” take-up of results in
policy or decision making? With the managers and decision makers, for not taking the trouble
to make themselves aware of new scientific developments? With the media, who are generally
content with repeating conventional wisdoms and pseudo-science rather than checking
authenticity? With the scientific community, who have a vested interest in assuring that no
question is finally answered so that research funding can continue? With inadequacies in the
structure and linkages between and within large organizations, which hinder information
flows? With the systems of economic incentives and penalties, and of regulation, which
govern land management and production practices? Probably all of these are responsible, and
new mechanisms, methods and research may be needed to overcome these constraints.

[IC-70] It has been argued (Calder, 1999) that another difficulty faced in the task of
connecting science to policy is the mindset that has been generated in relation to some
environmental issues. The perception that forests are always necessarily “good” for the water
resources has become so deeply ingrained in our collective psyches that it is usually accepted
unthinkingly. The view is routinely reinforced by the media and is all-pervasive; it has
become enshrined in some of our most influential environmental policy documents. The
report by the United Nations Commission on Environment and Development (UNCED, 1992)
states:

17
 The impacts of loss and degradation of forests are in the form of soil erosion; loss
of biological diversity, damage to wildlife habitats and degradation of watershed
areas, deterioration of the quality of life and reduction of the options for
development.

[IC-71] These simplistic views, particularly as they imply the inevitable link between the
absence of forests and “degradation” of water resources, have created a mindset which not
only links "degradation" with “less forest” but rehabilitation and conservation with “more
forest.” This mindset has caused, and continues to cause, governments, development agencies
and UN Agencies to commit funds to forestation or reforestation programmes in the mistaken
belief that this is necessarily the best or only way to improve water resources. Clearly, there
are many valid reasons for forestation or reforestation programmes, but where the objectives
are to improve water resources, they are unlikely to be achieved.

[IC-72] An example would be the ODA-funded forestry programme in Sri Lanka where
ODA in the early 1990s initiated a large-scale forestry programme on the Mahaweli
catchments in Sri Lanka based on the "old paradigm" assumption that pine reforestation
would "regulate flows and reduce erosion" to the catchments feeding the Victoria reservoir
complex. It is now realized (Calder, 1992; Finlayson, 1998) that the pine forestation is serving
only to reduce both annual and seasonal flows. Even if planted at the highest altitudes, where
there is still some remaining indigenous cloud forest, at Horton Plains, recent research
indicates that forestation would give no net benefit to flows: the measured interception losses
from the forest exceed the enhanced cloudwater deposition. Furthermore, where the planting
is generally taking place, on old, abandoned tea plantations, on-site soil erosion has virtually
ceased following the re-growth of pattana grasslands. Forestry operations involved with
planting and road construction will almost certainly be increasing on-site erosion.
Understorey fires under the pines, a common occurrence in Sri Lanka, also leave the soil
exposed to splash-induced erosion from the forest canopy. The forestry project is therefore
having the opposite effect to that intended. Furthermore, Stocking (1996) claims that even
where there is bad on-site erosion on the Mahaweli catchment, and erosion from tobacco
fields planted on steep slopes is probably the worst source of erosion on the catchment, this is
not getting into the reservoirs. Recent sedimentation studies reveal that there is, in fact, very
little sedimentation occurring in the reservoirs. Stocking claims that the on-site erosion from
the slopes is being deposited on the lower slopes and flood plains, and paddy field farmers are
actually benefiting from the sediment by incorporating it into their paddies.

4 The need for consistent policies

[IC-73] The need is surely undeniable for land use and water management policies that are
based on our best science rather than myth, and that are not only upwardly and downwardly
compatible, from the local rural watershed scale to the global scale, but are also consistent
with other global and local policies relating to sustainability, climate change, biodiversity,
trade, food production and poverty alleviation.

[IC-74] Perhaps nowhere is this need more acute than in relation to forests and water
policies. As discussed above, commercial forestry has often been promoted by development

18
organizations on its perceived environmental benefits. Yet science-based research has shown,
and shown for a long time, that many of the expected environmental benefits (which may in
some cases be provided by natural forests) cannot be achieved through commercial
plantations. Increasingly, we are now becoming aware of the environmental dangers, rather
than benefits, associated with such plantations. Not only is there usually a high cost in terms
of lost water associated with fast growing commercial plantations but, as has been recognized
by the government of South Africa, there may also be dangers associated with “escaping”
plantation trees. The Government of South Africa, in the February 2000 budget, awarded a
further rand 1 000 000 000 (over five years) to the Working for Water Programme (DWAF,
1996) for the purposes of controlling and eradicating alien invading tree species. The
expectation is that without this programme the invaders would eliminate indigenous plant
species and seriously reduce water resources. The programme also has, through specifically
targeting the poorest in society for employment, a major poverty alleviation component.
However, the expectation is that the water resource and ecological values of the programme
alone will justify the costs.

[IC-75] Although the alien invader issue has been recognized in South Africa, the extent of
this problem has not really been quantified for the rest of the developed and developing
world. The extent to which past and present policies of development organizations may have
been responsible for introducing and aggravating this problem – and are now morally
responsible for helping with its resolution – is perhaps another question that needs to be
broached.

[IC-76] As a means towards developing improved and more consistent land use and water
resource policies, applicable particularly at the rural watershed scale in developing countries,
the following questions are raised for discussion regarding the adequacy of present
approaches.

4.1 Compatibility of IWRM and SL approaches

[IC-77] Are Integrated Water Resources Management (IWRM) (resource focused) and
Sustainable Livelihoods (SL) (people focused) approaches necessarily complementary or
even compatible at all spatial and temporal scales?

[IC-78] The DFID (1997) White Paper on International Development presents the concept of
the stewardship of natural resources so that the needs of both present and future generations
can be met. The White Paper also promotes the concept of SL and this, together with the
management of “the natural and physical environment” is expected to achieve the overall goal
of poverty alleviation.

[IC-79] DFID provides the following definition for a sustainable livelihood:

A livelihood comprises the capabilities, assets (including both material and social
resources) and activities required for a means of living. A livelihood is sustainable
when it can cope with and recover from stresses and shocks and maintain or
enhance its capabilities and assets both now and in the future, while not
undermining the natural resource base.

19
[IC-80] Five types of assets, upon which individuals draw to build their livelihoods, are
generally defined: natural capital, social capital, human capital, physical capital and financial
capital (Carney, 1998).

[IC-81] The SL approach is still relatively new and untested and the consequences for natural
resource management of applying this new “people-first,” poverty-focused approach to
sustainable development is not yet known. There are concerns that poverty alleviation
approaches focused at the micro-catchment scale may result in “tragedy of the commons”
type impacts at larger scales. Conversely, it could be argued that IWRM approaches, although
aiming to achieve net economic benefits for basin inhabitants, might not be taking sufficient
account of the poorest in society.

4.2 Socio-economic impact of demand-management approaches

[IC-82] Do we fully understand the socio-economic implications of applying IWRM,

demand-management, economic instruments such as the registration and licensing of high-
water-use land uses (e.g. forests and sugar cane), water pricing and tradable water rights? Is
there a danger that, unless properly conceived, the poorest sectors of society will be
disadvantaged? Here again, there may be a conflict between IWRM approaches and poverty
alleviation; in particular the need to set water pricing at a level that the poorest in society can
afford. The South African Department of Water Affairs and Forestry (DWAF) now
recognizes that the primary goal of the Government of South Africa is the alleviation of
poverty. Barbara Schreiner, Chief Director, Water Use and Conservation, DWAF, explains
the issue eloquently (Schreiner, 1999):

There are those for whom water resources management appears to be an end in
itself, and for whom the challenge is to perfect the science, perhaps even the art,
of water resources management. For these people, economic growth is necessary
in order to provide the necessary context for improved water resources
management, as though improved water resources management has some inherent
and ultimate value in and of itself. Then there are those for whom water resources
management is a tool, and not an end. There are those for whom water resources
management is one element in the struggle to build a socially and environmentally
just society, not only in South Africa but across the whole world. For such people,
and I number myself amongst them, the ultimate purpose of what we do is to
create a society in which there is no more poverty, to create a world in which all
human beings have sufficient food and water, a place to live, a job, a clean and
healthy environment, education, and a chance for a life of dignity and self
fulfilment.

4.3 Carbon sequestration benefits and water resource dis-benefits of forests

[IC-83] Will the promotion of forestry as a land use in rural watersheds through carbon
sequestration credits (Kyoto Protocol) reduce downstream water resources and aggravate
downstream water resource conflicts?

20
[IC-84] One of the difficulties in reconciling forest policy in relation to both water and
climate change is that there are two separate research communities addressing these issues.
An example of this is given by the physical separation of the two communities into parallel
sessions at the Forests and Atmosphere-Water-Soil Conference held at Soltau, Northern
Germany, in July 1999.

[IC-85] The recognition that forest carbon sequestration benefits may be at the expense of
reduced water resources does not seem to have registered strongly in either community. Yet,
even if the principles underlying the Kyoto Protocol are accepted, there are important
questions to be asked by countries seeking carbon sequestration credits through forestation
policies, in relation to the value of these credits as compared with the value of the water that
will be forgone. In monetary terms they may be quite similar. Taking (conservative) estimates
of 250 and 400 mm per year as the average loss of water under deciduous hardwood and
pines/eucalypts respectively (Bosch and Hewlett, 1982) and (conservative) average values for
water for agricultural or industrial water use as US$ 100 per 1000 m3 (values have been
quoted as high as US$ 4000 for industrial processes and nearly US$ 2000 for some speciality
crops) the loss in value of water under hardwood forestation would be US$ 250/ha and
US$ 400 under pines/eucalypts. Aylward (Aylward, 1998; Aylward et al., 1998) suggests,
from studies in Latin America, that carbon sequestration credits arising from forestation
would amount to around US$ 200 to 500/ha, with a price of US$ 20/ton (amongst the highest
of the present carbon valuations) for carbon. Clearly, if there were higher-value uses for water
within a catchment and users were able to afford more than US$ 100 per 1000 m3 on strict
economic arguments alone, forgetting the possible socio-economic downsides, there would be
little merit in pursuing carbon credits.

4.4 Increased production efficiencies at the expense of downstream environmental

problems

[IC-86] The fallacy of striving for increased “irrigation efficiency” at the farm level, which
involves reducing the amount of “lost” return waters from the irrigation scheme, has been
ably pointed out by Seckler (see Seckler, 1996, and http://www.cgiar.org/iwmi//reps.htm). At
a larger, basin scale these return waters are not “lost” if they are being used for environmental
or other downstream purposes. Similarly, it should perhaps be argued that where increased
production efficiencies of other high-water-use land uses, whether irrigated farming, forestry
or agroforestry, are achieved at the expense of increased (water) resource use, the production
benefits need to be considered in relation to the marginal cost of the water to other
downstream users and, where necessary, compensation mechanisms to downstream users
should be considered.

[IC-87] For agroforestry systems we seem to have the particularly curious situation that,
irrespective of the downstream water consequences of increased productivity, increased
productivity from agroforestry systems as compared with separate forestry and agriculture,
seems impossible to achieve. Research has established that through competition for resources,
particularly water, the best that can normally be achieved from agroforestry systems is the
neutral productivity condition (i.e. the presence of trees will, for all known tree and
agricultural crop mixtures, reduce the productivity of the agricultural crop). Yet it would

21
appear that agroforestry is still being promoted by UN and other development agencies as a
means of increasing production, notwithstanding the extra labour requirement that
agroforestry systems usually entail. Is there a mismatch between the knowledge base and
policy in relation to the promotion of agroforestry as a means of increasing productivity?

4.5 Water subsidies to agriculture and increased downstream environmental problems

[IC-88] In most countries, government policies are such that irrigation water is still provided,
through subsidies, at an artificially low price. Increasingly, it is being recognized that the use
of these subsidized and low prices leads to inefficiencies in water use, which are direct
contributory factors to both waterlogging and downstream salinity problems.

[IC-89] Increasing the price for irrigation water is one method that has been adopted for
encouraging efficient water use. Another method, which has been used in the USA and
Australia, is the Transferable Water Entitlement (TWE) or transferable water right. This is a
mechanism by which a market for water can be achieved by allowing entitlements to be
bought and sold without the necessity of buying and selling the accompanying land. The use
of the mechanism is expected to increase efficiencies in a number of ways, including the
transfer of water to higher-value uses and higher-valued crops. It is also expected to lead to
the increased adoption of water-saving irrigation technologies, because the saved water can
then be sold. Decreased use of irrigation on land that is poorly suited and where economic
returns are low, perhaps because of existing waterlogged or salinized conditions, might also
be anticipated.

[IC-90] Arguably combinations of three approaches: land management, engineering solutions

and economic instruments, may hold the best prospects for the management of salinity and
downstream environmental problems

4.6 Matching local land use and water-resource policies with global ideals

[IC-91] The Global Water Partnership (GWP) document Towards Water Security,
Framework for Action (http://www.gwpforum.org/Vision.htm) presents again the messages
contained in the Dublin Statement and UNCED (UNCED, 1992) report of almost a decade
earlier. Some have questioned the lack of progress (e.g. “NGOs express serious concerns”
http://www.worldwaterforum.org/news/frameset2.cfm ) that has been made over this decade
and most would agree that further development of the ideals, and the huge costs that this
entails, is not what is required. What is required is action in terms of realistic policies
applicable at the local level. Where this is happening, and the case of South Africa, with a
new water act, and with the development of new sectoral strategies being developed, may be
the best example, perhaps we should be looking for “Best Management Practices,” to see how
such policies are performing and to see how they may be applied, with any necessary
modifications, elsewhere in the world?

[IC-92] It may also be necessary to explore the GWP proposals for consistency and to ask
questions such as “are some of these Global approaches useful or meaningful at the local,
rural catchment scale?”, e.g:

22
1. Is the GWP sending out an ambiguous message in calling for “Achieving water-food
security,” implying notions of national water-food security and self sufficiency, when
“virtual” water transfers through food trade may be a more efficient alternative?

2. A 30% increase in water productivity for food production from rainfed and irrigated
farming by 2015 is called for in the GWP document to meet future global food
requirements. Are such projections based on an adequate knowledge base? Are such
exhortations in any way meaningful at a local scale? Do they mask other economic and
socio-political realities currently constraining food production and water availability at the
rural catchment scale?

[IC-93] If it is agreed, during the conference, that some of the above questions are valid and
require an answer, perhaps a follow-up question that should be posed is “How can a solution
be found, what more must we do to ensure consistent, integrated policies within watersheds of
different scales that deal with land use impacts on water resources, with the impacts of
upstream activities on downstream users, that make use of appropriate economic, social and
political mechanisms within watersheds and that are making best use of current scientific
knowledge?".

Acknowledgements

The author wishes to thank all those that contributed to the Internet debate that was used here,
and to Thomas Hofer for his valuable editorial comments on this paper.

REFERENCES
Aylward, B. 1998. Beneficios y Costos De Oportunidad De La Conservación De Biodiversidad En El
Corredor Biológico Panameño - Componente Atlántico. Final Report to the Project on
Rural Poverty and Natural Resources. Panamá: INRENARE/MIDA.
Aylward, B., Echeverría, J., Fernández González, A., Porras, I., Allen, K., & Mejías, R. 1998.
Economic Incentives for Watershed Protection: A Case Study of Lake Arenal, Costa
Rica. p. 130, in: Final Report to the Government of the Netherlands under the program of
Collaborative Research in the Economics of Environment and Development (CREED).
London: IIED, TSC and the International Center for Economic Policy, National
University at Heredia (CINPE).
Bands, D.P., Bosch, J.M., Lamb, A.J., Richardson, D.M., Van Wilgen, B.W., Van Wyk, D.B., &
Versfeld, D.B. 1987. Jonkershoek Forestry Research Centre Pamphlet 384, Department
of Environment Affairs, Private Bag X447 Pretoria 0001. ISBN 0-621-11242-9.
Barr, J.J.F. 1998. Use of indigenous knowledge by natural resources scientists: issues in theory and
practice. Paper presented at the National Workshop on The State of Indigenous
Knowledge in Bangladesh. Dhaka, 6-7 May 1998.
Bosch, J.M. 1979. Treatment effects on annual and dry period stream flow at Cathedral Peak. S. Afr.
For. J., 108: 29-38.

23
Bosch, J.M., & Hewlett, J.D. 1982. A review of catchment experiments to determine the effects of
vegetation changes on water yield and evapotranspiration. J. Hydrol., 55: 3-23.
Bruijnzeel, L.A., & Bremmer, C.N. 1989. Highland-lowland interactions in the Ganges-Brahmaputra
river basin: A review of published literature. ICIMOD Occasional Paper, No.11.
Bruijnzeel, L.A. 1990. Hydrology of moist tropical forests and effects of conversion: a state of
knowledge review. UNESCO International Hydrological Programme, A publication of
the Humid Tropics Programme, UNESCO, Paris.
Calder, I.R. 1992. The hydrological impact of land use change (with special reference to afforestation
and deforestation). p. 91-101, in: Proceedings of the Conference on Priorities for Water
Resources Allocation and Management. Southampton, July 1992. Overseas Development
Administration, London. ISBN: 090 2500 49X.
Calder, I.R. 1998. Review outline of water resource and land use issues. (IIMI) SWIM Paper, No.3.
Calder, I.R. 1999. The Blue Revolution, Land Use and Integrated Water Resources Management,
Earthscan.
Cannell, M.G.R., Mobbs, D.C., & Lawson, G.J. 1998. Complementarity of light and water use in
tropical agroforests II. Modelled theoretical tree production and potential crop yield in
arid to humid climates. For. Ecol. Manag., 102: 275-282.
Carney, D. 1998. Implementing the Sustainable Rural Livelihoods Approach. In: D. Carney (ed)
Sustainable Rural Livelihoods, What contribution can we make? Papers presented at the
Department for International Development’s Natural Resources Advisers’ Conference,
July 1998. DFID, London. ISBN 1 86192 082 2.
CEASPA. 1997. Panama: Evaluation of National Sustainability. Panama Today, 7: 111
Chomitz, M., & Kumari, K. 1998. The domestic benefits of tropical forests: a critical review. The
World Bank Research Observer, 13(1): 13-35.
DFID [Department for International Development]. 1997. White Paper on International
Development. London.
DOE [Department of the Environment]. 1995. Rural England: a nation committed to a living
countryside ISBN: 0 10 130162 6. London: HMSO.
DWAF. 1996. The Working for Water Programme. Ministry of Water Affairs and Forestry, Cape
Town.
Finlayson, W. 1998. Effects of deforestation and of tree planting on the hydrology of the Upper
Mahaweli Catchment. Mahaweli Authority of Sri Lanka. ISBN 955 9185 02 0
Galay, V. 1985. Hindu-Kush Himalayan erosion and sedimentation in relation to dams. Paper
presented at the International Workshop on Watershed Management in the Hindu Kush-
Himalayan Region. Chengdu. ICIMOD, Kathmandu. 26 p.
Giambelluca, T.W., Fox, J., Yarnasarn, S., Onibutr, P., & Nullet, M.A. 1999. Dry-season radiation
balance of land covers replacing forest in northern Thailand. Agric. For. Meteorol.,
95(1): 53- 65.
Gowing, J.W., & Wyesure, G.C.L. 1992. Dry drainage, a sustainable and cost effective solution to
waterlogging and salinisation? Proceedings of the 5th International Drainage Workshop,
Lahore, Vol. III 6.26-6.34.
Hall, R.L., & Calder, I.R. 1993. Drop size modification by forest canopies - measurements using a
disdrometer. J. Geophys. Res., 90: 465-470.

24
Hamilton, L.S. 1987. What are the impacts of deforestation in the Himalayas on the Ganges-
Brahmaputra lowlands and delta? Relations between assumptions and facts. Mountain
Research and Development, 7: 256-263.
Hewlett, J.D., & Bosch, J.M. 1984. The dependence of storm flows on rainfall intensity and vegetal
cover in South Africa. J. Hydrol., 75: 365-381.
Hewlett, J.D., & Helvey, J.D. 1970. Effects of forest clearfelling on the storm hydrograph. Water
Resour. Res., 6(3): 768-782.
Hofer, T. 1998a. Floods in Bangladesh. A highland-lowland interaction? Geographica
Bernensia G 48.
Hofer, T. 1998b. Do land use changes in the Himalayas affect downstream flooding? Traditional
understanding and new evidences. Memoir Geological Society of India 19: 119-141.
Ives, J.D., & Messerli, B. 1989. The Himalayan dilemma. Reconciling development and
conservation. London: United Nations University Press.
Johnson, R.C. 1995. Effects of upland afforestation on water resources: the Balquhidder experiment
1981-1991. Institute of Hydrology (Wallingford) Report, No.116. 73 p.
Jones, J.A., & Grant, G.E. 1996. Peak flow responses to clear-cutting and roads in small and large
basins, western Cascades, Oregon. Water Resour. Res., 32: 959-974.
Kirby, C., Newson, M.D., Gilman, K. 1991. Plynlimon research: the first two decades. Institute of
(Wallingford) Hydrology Report, No.109. 188 p.
Langford, K.J. 1976. Change in yield of water following a bushfire in a forest of Eucalyptus regnans.
J. Hydrol., 29: 87-114.
Lauterburg, A. 1993. The Himalayan highland-lowland interactive system: do land use changes in the
mountains affect the plains? In: B. Messerli, T. Hofer and S. Wymann (eds) Himalayan
environment: pressure-problems-processes. 12 years of research. Geographica Bernensia
G38, University of Berne.
Ong, C.K., Odango, J.C.W., Marshall, F., & Black, C.R. 1991. Water use by trees and crops. Five
hypotheses. Agroforestry Today, April-June: 7-10.
Ong, C.K., Black, C.R., Marshall, F., & Corlett, J.E. 1996. Principles of resource capture and
utilization of light and water. p. 73-158, in: C.K. Ong, and P. Huxley (eds) Tree-Crop
Interactions. CAB International, Wallingford, UK. ISBN 0-85198-987-X.
Ong, C.K. 1996. A framework for quantifying the various effects of tree-crop interactions p. 1-23, in:
C.K. Ong, and P. Huxley (eds) Tree-Crop Interactions. CAB International, Wallingford,
UK. ISBN 0-85198-987-X.
Pearce, A.J. 1986. Erosion and sedimentation. Working paper. Environment and Policy Institute,
Honolulu, Hawaii. 18 p.
Pereira, H.C. 1989. Policy and practice in the management of tropical watersheds. Westview Press,
Colorado. ISBN 0 8133 7731 5.
Ranganathan, R., & de Wit, C.T. 1996. Mixed cropping of annuals and woody perennials: an
analytical approach to productivity and management. p. 25-49, in: C.K. Ong, and P.
Huxley (eds) Tree-Crop Interactions. CAB International, Wallingford, UK. ISBN 0-
85198-987-X.
Rao, M.R., Sharma, M., and Ong, C.K. 1990. A study of the potential of hedgerow intercropping in
semiarid India using a two-way systematic design. Agrofor. Syst., 11: 243-258.

25
Reid, I., & Parkinson, R.J. 1984. The nature of the tile-drain outfall hydrograph in heavy clay soils. J.
Hydrol., 72: 289-305.
Robertson, G. 1996. Saline Lands in Australia its extent and predicted trends. Proceedings of the 4th
National Conference and Workshop on the Productive Use and Rehabilitation of Saline
Lands. Published by Promaco Conventions PTY Ltd, Australia.
Robinson, M., Moore, R.E., & Blackie, J.R. 1997. From Moorland to Forest: The Coalburn
Catchment Experiment, Institute of Hydrology and Environment Agency Report,
Institute of Hydrology, Wallingford, UK.
Robinson, M., Ryder, E.L., & Ward, R.C. 1985. Influence on stream flow of field drainage in a small
agricultural catchment. Agricultural Water Management, 10: 145-158.
Scott, D.F., & Lesch, W. 1997. Stream flow responses to afforestation with Eucalyptus grandis and
Pinus patula and to felling in the Mokobulaan experimental catchments, South Africa.
J. Hydrol., 199: 360-377.
Scott, D.F., & Smith, R.E. 1997. Preliminary empirical models to predict reduction in total and low
flows resulting from afforestation. Water SA, 23: 135-140.
Schreiner, B. 1999. Keynote address: SANCIAHS Conference, Cape Town, 29 November 1999,
Department of Water Affairs and Forestry, South Africa.
Seckler, D. 1996. The new era of water resources management. IIMI Research Report, No.1.
Stocking, M. 1996. Soil erosion - breaking new ground. p. 140-154, in: M. Leach and R. Mearns
(eds) The Lie of the Land. London: Villiers.
Taylor, C.H., & Pearce, A.J. 1982. Storm runoff processes and sub-catchments characteristics in a
New Zealand hill country catchment. Earth Surf. Process. Land forms, 7: 439-447.
UNCED [United Nations Conference on Environment and Development]. 1992. Agenda 21 & the
UNCED proceedings. Proceedings of the UNCED Conference, Rio de Janeiro, Brazil.
1992. UNCED, New York, ISBN 0379103508
USDA (U.S. Department of Agricultural) Agricultural Research Service. 1961. A Universal equation
for Predicting Rainfall-erosion Losses. USDA-ARS Spec. report. 22-26.
van Lill, W.S., Kruger, F.J., & van Wyk, D.B. 1980. The effects of afforestation with Eucalyptus
grandis (Hill ex Maiden) and Pinus patula (Schlecht. Et Cham.) on stream flow from
experimental catchments at Mokobulaan, Transvaal. J. Hydrol,, 48: 107-118.
van Noordwijk, M., Lawson, G., Soumaré, A., Groot, J.J.R., & Hairiah, K. 1996. Root distribution of
trees and crops: competition and/or complementarity. P. 319-364, in: : C.K. Ong, and P.
Huxley (eds) Tree-Crop Interactions. CAB International, Wallingford, UK. ISBN 0-
85198-987-X.
Versfeld, D.B. 1981. Overland flow on small plots at the Jonkershoek Forestry Research Station.
South African Forestry Journal, 119: 6.
Zhang, L., Hume, I.H., O’Connell, M.G., Mitchell, D.C., Milthorpe, P.L., Yee, M., Dawes, W.R., &
Hatton, T.J. 1999. Estimating episodic recharge under different crop/pasture rotations in
the Mallee region. Part 1. Experiments and model calibration Agric. Water. Manag.,
42(2): 219-235.

26