Groundwater discharge as affected by land use change in small
catchments: A hydrologic and economic case study in Central Brazil1
Henrique M. L.Chaves[1], Ana Paula S. Camelo[2], & Rejane M. Mendes[3]
[1] Professor, School of Technology-EFL, University of Brasilia-UnB, 70910-900, Brasilia-DF Brazil, E-mail:
[email protected] (corresponding author)
[2] Graduate assistant, School of Technology-ENC, University of Brasilia-UnB, 70910-900, Brasilia-DF Brazil.
[3] Forest engineer and consultant. Formerly: undergraduate assistant, School of Technology-EFL, University
of Brasilia-UnB, 70910-900, Brasilia-DF Brazil.
Abstract. The Pipiripau river basin, a 235 km2 catchment in central Brazil, has experienced a significant
increase in land use intensity in the last 40 years. This has contributed to a significant reduction in the base flow
discharge, responsible for the maintenance of the stream flow during the dry winter season. To assess the
hydrological and economic benefits of three land conservation programs in the basin, an empirical correlation
was obtained between the base flow index and the normalized basin curve-number, calibrated with observed
stream flow and precipitation data. The results indicate that if reforestation and best management practices are
implemented in the basin, up to 755 x 106 m3yr-1 of additional base flow discharge would result during the dry
season, with additional revenues up to US$ 1.03 million yr-1 for the water utility company. The simplicity of the
presented methodology allows for its application to other data-scarce tropical basins.
1. Introduction
Base flow discharge is an important element of the hydrological cycle that describes the loss of
water from the groundwater compartment to surface waters (GRAPHIC/UNESCO, 2006). Often
equaled to groundwater discharge in river basins, base flow discharge helps to maintain aquatic
ecosystems and provides for important hydrological services to human communities (Tomish et al.,
2004).
Groundwater discharge is influenced by climate, watershed, and land use/management conditions.
Forests are an important component in the stabilization of groundwater discharge and stream flow
(Pereira, 1989). When forest conversion is followed by land uses that interfere and alter natural bio-
physical processes, hydrological consequences may follow (Pattanayak, 2004). Although annual and
storm flows typically increase with forest conversion to agriculture, base flows often decline owing to
reduced infiltration and more episodic export of water (Poff et al., 1997; Siriwardena et al., 2005).
Several studies have highlighted the impacts of land use changes to groundwater discharge in
river basins. Sloto (2008) reported that a decrease of up to 63% of base flow discharge would occur
following full build-out of a small catchment in Pennsylvania. Klöcking & Haberlandt (2002),
analyzing the effect of forestation and deforestation scenarios in base flow discharge at the outlet of
five German catchments, reported that base flow discharge would vary from +3% to -15%, with
respect to the baseline conditions.
In general, best management practices are known to improve base flow discharge in river basins.
Schilling & Libra (2003) reported that annual base flow percentage increased between 20 and 30% in
two catchments in Iowa in a 60-year period, following soil and water conservation projects during the
1940’s.
Basin scale can also affect the response of land use change in groundwater discharge studies. In
small basins, the influence of land use is more significant, whereas geologic and climatic factors
become more prevalent in large catchments (Pattanayak, 2004).
Although data intensive, process-based mathematical models are capable of reproducing the
surface and groundwater processes holistically and accurately in river basins, the lack of information
often hinders their application in the tropics (Chaves, 1996).
1
Please do not circulate.
1
In recent years, economic valuation of environmental services and their corresponding payments
are becoming more and more effective in the solution of water quality and quantity issues in river
basins (Classen et al., 2001, Chaves et al., 2004). Pattanayak (2004) has shown that reforestation
programs in small Indonesian catchments would increase the percentage of base flow discharge up to
24%, generating an equivalent amount of savings in water collection costs in small communities.
1.1. Purpose and Scope
The purpose and scope of the present work was the analysis of the hydrological and economic
impacts of land use change in a small basin in Brazil, particularly regarding base flow discharge
responses. The study follows the UNESCO/GRAPHIC philosophy, namely the assessment of
groundwater resources under the pressures of humanity and climate changes. In addition to the
hydrologic response, economic valuation is important because its results are more easily
comprehended by water users and decision-makers.
Additionally, economic valuation of hydrologic services allow for the establishment of incentive
payment programs, which help the mitigation of exiting hydrologic problems, as well as the
adaptation to future impacts.
1.2. Study Area Description
The area studied was the Pipiripau river basin, situated in the northeastern corner of the Federal
District, in central Brazil (Figure 1).
Figure 1. Location of the Pipiripau river basin, with respect to the Federal District area (Brazil).
The Pipiripau river basin has an area of 235 km2, with central coordinates 15o27’14”S and
47 27’47”W. The river basin has a mean altitude of 950 m, gentle topography (average slope of
o
5.5%), and deep, well drained soils (red oxisols & ultisols), underlain by quartzites, phyllites, and
rhythmites, the latter presenting medium to low permeabilities and transmissivities (Chaves & Piau,
2008).
Presently, the main land uses in the catchment are agriculture (43% of the basin area),
pastureland (28%), and natural savannah (29%). Figure 2 shows the main land uses in the basin at the
present.
2
Figure 2. Present land uses in the Pipiripau river basin. Source: Chaves & Piau (2008).
Mean annual precipitation in the basin is 1,300 mm, with wet summers and dry winters. Mean
annual temperature is 22oC. Long-term mean annual stream flow, measured at the basin outlet, is 2.9
m3s-1, and base flow discharge represents 85% of the yearly mean stream flow (Chaves & Piau, 2008).
1.3. Relevance of the Case Study with respect to GRAPHIC issues
In the last 20 years, the main water uses (urban water supply and irrigation) have been competing
for the river water in the basin, particularly during the long dry winters. However, both users have to
observe their water allocation limits, in addition to the minimum restriction (environmental) flows,
established by federal and state water agencies.
Since the river is not regulated by dams, the limiting water supply during the winter months often
lead to water shortages, hindering the development of new businesses and other water-dependent
activities.
Additionally, base flow discharges, responsible for maintaining the stream flow of the Pipiripau
river during the dry season, has been decreasing in the last 40 years, leading to further water-use
conflicts in the basin. Figure 3 presents the mean dry-season stream flow values (mean stream flow in
the driest 6 months), corresponding to the base flow discharge in the period.
3
4,0
3,0
Qd
(m3s-1) 2,0
1,0
-
1970 1980 1990 2000 2010
Year
Figure 3. Time series of the mean yearly dry season stream flow in the Pipiripau river.
The dashed line in the time series of Figure 3 shows a significant (95% probability) decreasing
linear trend in the dry season discharge in the Pipiripau river. Possible explanations for this behavior
are the decreasing trend of annual precipitation in the period, reduced groundwater recharge in the
basin as a consequence of increasing land development, increasing water abstractions from the river,
or a combination of the three.
The first process has to do with climate variability / change, and the latter two with human
pressures in the basin, in the recent past. It is likely that the climate and human pressures are nested,
making it difficult to pinpoint the leading cause for the base flow reduction, and therefore to establish
appropriate mitigation and adaptation policies.
1.4 Contribution of the Case-study to GRAPHIC long-term goals
Since the objectives of the present work were to correlate the decreasing groundwater discharge
trends to the basin land use intensity, to establish a relationship between base flow discharge and
basin hydrologic parameters, and to devise an empirical method that could estimate hydrological
services resulting from land conservation practices in the basin, it is expected that the results could
contribute to the GRAPHIC subjects, methods, and experiences.
Additionally, the methodology utilized in the present case study could be applied in similar
areas, particularly in data-scarce tropical basins.
2. Methodology
In order to achieve its objectives, the methodology of the present case-study was divided in four
steps: i) correlation the annual base flow discharge with the basin land use intensity in the last 40
years; ii) determination of basin curve-number and base flow discharge from existing precipitation
and stream flow records; iii) obtaining an empirical relationship between the base flow index, land
use change and climate variability; and iv) using the obtained relationship to estimate the hydrological
services and economic benefits resulting from three land conservation scenarios in the basin.
2.1. Correlating annual base flow discharge with basin land use intensity
In this step, basin land use dynamics in the last three decades was obtained from temporal
analysis of land-use maps, generated from Landsat satellite images (1984-2006). For that purpose, a
multi-band, maximum likelihood supervised classification routine (Mather, 1999) was used.
4
The proportions and frequencies of the land use classes in the basin were obtained for five
periods: 1984, 1994, 1998, 2001, and 2006. Additionally, a land use intensity routine was performed
in the classes and frequencies of the resulting land use maps. The land-use intensity index used was
(Ometo et al., 2000; Chaves & Santos, 2009):
∑ [1]
Where: LUI = land use intensity index, fi = proportion of the land class i in the map, and wi =
weight of the land use class i. The weights in equation [1] increase as the land use intensity increases,
i.e., w = 1 for natural conditions, w = 2 for little disturbances in the natural condition, w = 3 for
agriculture and pasture, w = 4 for low density urban areas, and w = 5 for dense urban areas (Chaves &
Santos, 2009).
In order to avoid climate interference, mean annual stream flow of the driest 6 months (Qd) of the
Pipiripau river was normalized by annual basin precipitation. The normalized variable Qd/P was then
plotted against the corresponding LUI for the five years studied, and an empirical relationship was
obtained.
2.2 Obtaining the basin curve-number and base flow discharge from stream flow data
The next step was the estimation of the basin curve-number coefficient (CN2) from the daily
rainfall and stream flow records for the 1991-2009 period, where land use conditions in the basin
were relatively stable. In order to separate storm runoff from base flow, the daily stream flow data
was passed through a digital filter, developed by Nathan & MacMahon (1990):
- ( - ) [2]
Where: qt (m3 s-1) = filtered stream flow; Qt (m3 s-1) = original stream flow at day i; and α = filter
parameter, taken as 0.925 (Nathan & MacMahon, 1990). Daily base flow discharge (Qb) is simply the
difference between Qt and qt in equation [2].
After the hydrograph separation, significant storm flow hydrographs, resulting from precipitation
events greater than 10 mm in three consecutive days, were integrated, and a runoff volume (Qr, in
mm) was obtained. A corresponding CN2 value for each storm hydrograph was obtained through the
iterative solution of the following equation (NRCS, 1972):
( )
[3]
( )
Where: Qr (mm) = runoff volume of the storm flow hydrograph, P (mm) = precipitation of the event,
CN2 = curve number for medium antecedent moisture conditions. Appropriate corrections were made
in the calculated CN2 whenever drier (Pa< 3.56 mm in the previous 5 days) or wetter (Pa > 5.33 mm)
antecedent soil moisture conditions (AMC) occurred (Chow et al., 1988):
(dry AMC) or (wet AMC) [4]
Annual means of calibrated CN2 values were obtained for each hydrologic year in the series, and
a global mean of CN2 obtained for the basin. To verify the accuracy of the calibrated CN2 value, a
weighted mean CN2 value was obtained from the NRCS (1972) table, with the aid of the GIS and
basin land use maps.
5
2.3 Empirical relationship between the base flow index and the normalized runoff coefficient
In this step, the dimensionless yearly mean base flow index (Qb/Q) of the basin was plotted
against the normalized yearly curve number (CN2/P), in the period 1992-2009. Outliers with too high
or too low base flow indices were removed from the analysis. The obtained empirical relationship was
later used in the estimation of the hydrologic services resulting from different land conservation
scenarios, by corresponding changes in the CN2 coefficient.
2.4. Estimating and valuing hydrological services resulting from land conservation scenarios
Since reforestation and best management practices (BMPs) can increase base flow discharge
(Klöcking & Haberlandt, 2002; Schilling & Libra, 2003), the mean annual base flow increment in the
Pipiripau river basin was estimated under three hypothetic land use scenarios, starting from the
present (baseline) land use conditions.
Land conservation scenarios were defined based on the environmental liabilities previously
identified in the basin, including forest cover deficits in riparian strips and property reserves
(reforestation), and unprotected aquifer recharge areas (BMPs).
Table 1 below lists the three land conservation scenarios analyzed in the study. These scenarios
considered three possible climate conditions: average (mean annual precipitation), dry (mean annual P
minus one standard deviation), and wet (mean annual P plus one standard deviation) years.
Table 1. Land conservation scenarios and climate conditions used in the estimation of
the hydrologic services in the Pipiripau river basin.
Scenario % of Climate
Type of land conservation practice
Number Basin Area condition
Reforestation of riparian and legal reserve areas 9.5%
1 D/A/W
with native tree species
BMPs in agricultural and pastureland in 44.1%
2 D/A/W
recharge areas (no-till agriculture, terraces)
3 Both scenarios (1+2) 53.6% D/A/W
According to Table 1, a total of nine combinations of land use/climate scenarios were
analyzed. In order to estimate the mean annual base flow discharge increment resulting from the each
of the nine scenarios above, and starting from the calibrated CN2 value for the basin, new weighted
CN2 values were obtained with the aid of the GIS and Table 2 below.
Table 2. CN values for a combination of soil type and land use. Source: NRCS (1972).
Hydrologic Soil Group
Land Use Land Management
A B C D
Without conservation practices 72 81 88 91
Agriculture
With conservation practices 62 71 78 81
Poor condition 68 79 86 89
Pastureland
Good condition 39 61 74 80
Savannah Good condition 30 58 71 78
Regular cover 45 66 77 83
Woodland
Good cover 25 55 70 77
6
The new curve-number (CN2’) of the basin under the different land conservation scenarios was
the weighted mean of CN2 values of Table 2 and the land use / management frequencies:
[5]
Where: CN2’= the curve-number under the land conservation scenario, CN2i = curve-number under
land use/management i, Ai = ith area corresponding to the land-use combination i, and A = total basin
area.
The base flow discharge under different land use and climate scenarios was computed with an
empirical linear regression equation of the form:
Qb / Q = a (CN2 / P) + b [6]
Where: a = the angular coefficient of regression, b = the linear coefficient, Q and Qb are the stream
flow and base flow discharge (m3s-1), respectively, and P (mm) is the annual precipitation.
After Qb was calculated in the baseline land use and climate condition, a new Qb was computed
by equation [5] for the desired land use and climate combination. P & Q values in equation [6] were
set at three levels each (dry, average, and wet conditions). The increment in annual base flow
discharge for a given land conservation/climate condition was therefore:
[7]
Where: Qb (m3s-1) = mean annual increment in base flow discharge in the basin, Qb1 (m3s-1) = mean
annual flow discharge under the new land conservation/climate scenario, and Qb0 (m3s-1) = mean
annual flow discharge under the baseline scenario.
Once the base flow discharge increments were obtained for each land conservation/climate
combination, the resulting economic externality was estimated by the product of the annual volume of
the base flow increment and the average household water price in the basin (US$ 1.5 m-3).
3. Results and Discussion
3.1. Correlation of the dry season discharge and basin land use intensity
Figure 4 presents the land use dynamics in the basin between 1973 and 2006. According to that
Figure, savannah, woodland and grassland have been replaced with agriculture and other anthropic
uses in a steady fashion.
7
80
Agriculture & Pasture
60
Natural grassland
Savannah
Area Woodland
(%) 40 Exposed soil
Urban
Water
20 Eucalyptus
0
1973 1984 1994 1998 2001 2006
Figure 4. Land use dynamics in the Pipiripau river basin in the period 1973-2006.
Figure 5 below shows the correlation between land use intensity index (LUI, equation [1]) and
the normalized dry season discharge (Qd/P) in the Pipiripau river basin during the period 1984-2006.
0,30
0,25
0,20
Qd / P
0,15
0,10
0,05
0,00
1,5 2,0 2,5 3,0 3,5 4,0
LUI
Figure 5. Relationship between the normalized dry-season stream
flow and the land use intensity index in the Pipiripau river basin
between 1984 and 2006.
LUI almost doubled in the period studied, from 1.88 in 1984 to 3.43 in 2006, with a correspoding
decrease in Qd/P. A good fit (R2= 0.92) was obtained for the linear regression in Figure 5, regardless
of the rainfall variability in the period.
Mixed results relating land use conversion and base flow discharge were found in the literature.
While Krause (2002) reported a significant decrease of the base flow index (Qb/Q) as a result of
conversion of forests to agriculture in a German basin, Siriwardena et al., (2006) found the opposite
in a basin in Queensland (Australia). Considering that other factors (rainfall variability,
evapotranspiration) play a role in base flow discharge generation, they could have affected the
contrasting results.
3.2 Base flow discharge hydrographs and basin curve-number (baseline condition)
Table 3 below presents the yearly precipitation, stream flow (Q), base flow discharge (Qb), base
flow index (Qb/Q), and the calibrated CN2 values for the Pipiripau river basin, in the 1992-2009
period (baseline condition).
8
Table 3. Yearly precipitation (P), mean stream flow (Q), mean base flow
discharge (Qb), base flow index (Qb/Q), calibrated CN2, and normalized CN2
(CN2/P) for the Pipiripau river basin, in the period 1992-2009.
P Q Qb Qb/Q CN2 CN2/P
Year
(mm) (m3s-1) (m3s-1)
1992 1,417 4.0 3.5 0.877 73.4 0.052
1993 1,383 4.6 4.1 0.887 73.4 0.053
1994 1,254 3.2 2.7 0.857 79.2 0.063
1995 1,431 2.1 1.9 0.908 66.4 0.046
1996 951 1.4 1.2 0.855 69.6 0.073
1997 1,479 2.0 1.6 0.807 73.3 0.050
1998 1,085 2.0 1.7 0.841 81.6 0.075
1999 1,340 1.8 1.5 0.840 72.0 0.054
2000 1,063 2.0 1.6 0.823 73.9 0.069
2001 912 1.2 1.0 0.809 74.1 0.081
2002 1,123 1.7 1.3 0.784 73.0 0.065
2003 1,213 1.6 1.3 0.851 68.2 0.056
2004 1,627 3.4 2.7 0.795 75.3 0.046
2005 1,316 2.7 2.3 0.862 70.4 0.054
2006 1,428 3.0 2.6 0.868 69.8 0.049
2007 1,228 2.5 2.1 0.859 69.0 0.056
2008 1,158 1.5 1.4 0.883 68.1 0.059
2009 1,147 1.7 1.5 0.882 68.2 0.059
Means 1,253 2.4 2.0 0.849 72.2 0.059
The mean calibrated CN2 for the catchment in the baseline condition was 72.2, very close to the
estimated (weighted) average (CN2= 70.7), indicating that the hydrograph separation process was
adequate.
Mean yearly precipitation in the period was 1,253 mm, with a standard deviation of 191 mm.
Mean base flow was 2.0 m3/s, and mean base flow index was 0.85. Figure 6 presents the correlation
obtained between the base flow index and the normalized CN2 in the Pipiripau river basin.
0,92
0,90
y = -1.83 x + 0.97
R = -0.74
0,88
Qb / Q
0,86
0,84
0,82
0,80
0,040 0,050 0,060 0,070 0,080 0,090
CN2 / P
Figure 6. Correlation between the base flow index and the
normalized CN2 for the Pipiripau river basin during studied period.
Figure 6 shows that a relatively good fit was obtained for the linear regression between Qb/Q and
CN2/P. The dispersion in the data reflects the changes in watershed/aquifer conditions in the period
studied.
Figure 6 indicates that, for a given climate condition, an increase in CN2 would decrease the
base flow index, i.e., the relative contribution of base flow discharge to total stream flow. The
9
advantage of using the empirical function of Figure 6 is that the normalization in the dependent and
independent variables eliminate intrinsic climatic variability, allowing for the isolation of the land use
effects in base flow discharge.
Although there is a resemblance between Figures 5 and 6, the independent variable in the latter
(CN2/P) has more hydrologic meaning, and therefore it is more useful for predictions of future basin
behavior.
Figure 7 below explains part of the behavior of decreasing base flow index with increasing CN2,
as seen from Figure 6. The former shows that the number of days with direct runoff in the Pipiripau
basin has increased linearly in the period studied, contributing for the reduction of groundwater
recharge and base flow discharge.
160
120
Days with
runoff 80
40
Year
Figure 7. Number of days of the year with storm runoff in the Pipiripau river basin.
3.4. Hydrological services resulting from land conservation scenarios
The weighted average of CN2 for the Pipiripau river basin under the different land conservation
scenarios are presented in Figure 4.
Table 4. Calibrated (baseline) and prospective
(scenarios) CN2 values for the Pipiripau river basin.
Scenario Baseline Scenario Scenario Scenario
1 2 3
CN2 72.2 69.2 64.8 59.7
According to Table 4, there would be a significant reduction in the basin CN2 after the
introduction of the land conservation scenarios, the highest being under scenario 3, involving
reforestation and the implementation of best management practices in agricultural areas.
Figure 8 shows the hydrological services with respect to base flow discharge (delta Qb) and the
additional revenue for the water utility company (delta R) as a function of the three land conservation
scenarios, under the three climate conditions analyzed.
10
1200
1000
800
600 Delta Qb (1,000 M3/yr)
Delta R (1,000 US$/yr)
400
200
0
1D 1A 1W 2D 2A 2W 3D 3A 3W
Land conservation & climate scenarios
Figure 8. Hydrological services and additional revenues resulting
from the land conservation scenarios in the Pipiripau river basin.
According to Figure 8, there would be a substantial increase in the dry season water availability
in the basin, from 94,300 m3 in a dry year under scenario 1 to 754,800 m3 in a wet year under scenario
3. The additional revenues would range from US$ 123,900 yr-1 to US$ 1,032,400 yr-1, respectively.
An increase of 25% in the base flow index of nine American basins was reported by Schilling &
Libra (2003), as a result of improved conservation practices during the last 40 years, relative to
previous conditions.
Although the land conservation scenarios could potentially increase basin evapotranspiration
(ET) and decrease base flow discharge, the small proportion of reforestation (< 10% of the basin area)
of scenarios 1 and 3 would not affect ET in a significant manner. In the case of scenario 2, the mulch
cover of no-till agriculture would in fact reduce ET, with respect to the baseline conditions.
4. Policy-relevant Recommendations
The additional revenues generated by the increased base flow discharge (Figure 8) could be used
to pay for the hydrologic services provided by participating farmers, creating a virtuous circle of land
conservation and flow stabilization in the basin.
The hydrologic services and revenues would also show the decision makers and water managers
the economic feasibility of land conservation, which could lead to the development and
implementation of land conservation projects in the future. In addition to the mitigation of the land
use intensification in the basin, land conservation scenarios would also provide for an adaptation
strategy against future climate change.
Although there are supporting evidences of the positive effect of land conservation on base flow
discharge in the literature, the results of the present case-study reflect a specific situation, and could
not be directly transferred to other river basins. However, the simplicity and robustness of the
presented methodology facilitate its application to other data-scarce river basins, where data intensive,
process-based models could not be used.
5. Future Work
Considering that the Pipiripau river basin has been an object of research in the last ten years,
future activities in the Pipiripau river basin will involve:
The survey and mapping of groundwater levels in the basin;
The downscaling of GCM outputs to the basin scale and estimation of future climate change
impacts in basin hydrology;
The application of process-based models in the basin, such as SWAT;
11
Continuous monitoring of stream flow and base flow in the basin, during and after the
implementation of land conservation programs;
The analysis of the vulnerability and risk of groundwater contamination by pesticides;
Hydrologic and economic effectiveness of the implementation of agri-environmental
(incentive payment) programs in the basin;
The comparison of the results with other GRAPHIC case studies.
6. References
Chaves, H.M.L. 1996. Mathematical modeling of water erosion: past, present, and future. In: Alvarez,
V.H. (Ed.): The soil in the large morpho-climatic domains of Brazil. UFV press, Viçosa, p. 731-
750 (in Portuguese).
Chaves, H.M.L., B. Braga, A.F. Domingues, & D.G. dos Santos. 2004. Estimating the environmental
services and financial payments of ANA’s Water Provider Program. R. Bras. Rec. Hídricos 9(3):5-
14 (in Portuguese).
Chaves, H.M.L. & Piau, L.P. 2008. Effects the rainfall variability and of the land-use and
management on the direct runoff and sediment yield of a small catchment in the Federal District of
Brazil. Rev. Bras. Ci. Solo 33: 333-343 (in Portuguese).
Chaves, H.M.L. & L.B. Santos. 2009. Land use, landscape fragmentation, and water quality in a small
watershed. R. Bras. Eng. Agric. Amb. 13(supl.):922-930.
Chow, V.T., D.R. Maidment, & L.W. Mays. 1988. Applied hydrology. McGraw-Hill, N. York, 572 p.
Claassen, R., L. Hansen, M. Peters, V. Breneman, M. Weinberg, A. Cattaneo, P. Feather, D. Gasby,
D. Hellerstein, J. Hopkins, P. Johnson, M. Morehart, & M. Smith. 2001. Agri-environmental
policy at the crossroads: guideposts on a changing landscape. USDA-ERS Report No. 794,
Washington. 67 p.
GRAPHIC-UNESCO. 2006. Groundwater resources assessment under the pressures of humanity and
climate change. International Hydrologic Program/Unesco, Paris. 19 p.
Klöcking, B. & U. Haberlandt. 2002. Impact of land use changes on water dynamics – A case study in
temperate meso and macroscale river basins. Phys. & Chem. Earth 27: 619-629.
Krause, P. 2002. Quantifying the impact of land use changes on the water balance of large catchments
using the J2000 model. Phys. & Chem. Earth 27: 663-673.
Mather, P.M. 1999. Computer processing of remotely-sensed images – An introduction. Wiley, N.
York, 292 p.
Nathan, R.J. & T.A. McMahon. 1990. Evaluation of automated techniques for base flow and
recession analysis. Water Res. Res. 26(7):1465-1473.
Natural Resource Conservation Service. 1972. National engineering handbook – Chapter 4:
Hydrology. Washington.
Ometo, J., L.A. Martinelli, M.A. Ballerster, A. Gessner, A.V. Krusche, R.L. Victoria, & M. Williams.
2000. Effects of land use on water chemistry and macro-invertebrates in two streams of the
Piracicaba river basin, Brazil. Freshwater biol. 44:327-337.
Pattanayak, S.K. 2004. Valuing watershed services: Concepts and empirics from southeast Asia. Agr.
Ecos. & Env. 104: 171-184.
Pereira, H. 1989. Policy and practice in the management of tropical watersheds. Westview Press,
Boulder, 125 p.
Poff, N.L., D.J. Allan, M.B. Bain, J.R. Karr, & K.L. Prestegaard. 1997. The natural flow regime: a
paradigm for conservation and restoration of riverine ecosystems. Bioscience 47:769-784.
12
Schilling, K.E. & R.D. Libra. 2003. Increased base flow in Iowa over the second half of the 20th
century. J. Am. Water Res. Assoc. 39(4): 851-860.
Sloto, R.A. 2008. Effects of land-use changes and groundwater withdrawals on stream base flow,
Pocono creek watershed, Monroe County, Pennsylvania. Scientific Investigation Report No. 2008-
5030, USGS, Washington, 50 p.
Tomich, T.P., D.E. Thomas, & M. van Noordwijk. 2004. Environmental services and land use change
in southeast Asia: From recognition to regulation or reward? Agr. Ecos. & Env. 104: 229-244.
13
