Received: 4 August 2020 Revised: 20 October 2021 Accepted: 26 October 2021

DOI: 10.1002/ird.2662

RESEARCH ARTICLE

Impact of agricultural water management interventions on

upstream–downstream trade-offs in the upper Cauvery
catchment, southern India: a modelling study*

Pawan S. Wable1 | Kaushal K. Garg1 | Rajesh Nune1 |

Akuraju Venkataradha1 | Anantha KH1 | Veena Srinivasan2 | Ragab Ragab5 |
John Rowan4 | Virginie Keller5 | Pradeep Majumdar3 | Gwyn Rees5 |
Ramesh Singh1 | Sreenath Dixit1

1
ICRISAT Development Centre,
International Crops Research Institute for
Abstract
the Semi-Arid Tropics, Hyderabad, The Cauvery basin in southern India is experiencing transboundary issues
Telangana, India due to increasing water demand. This study analysed water balance compo-
2
Ashoka Trust for Research in Ecology
nents and the impact of agricultural water management (AWM) interven-
and the Environment (ATREE),
Bengaluru, India tions in the upper Cauvery catchment of the Cauvery basin. Results showed
3
Indian Institute of Science (IISc), that the study catchment receives an average of 1280 mm of annual rainfall.
Bengaluru, India Of this, 29% (370 mm) flows downstream, 54% (700 mm) contributes to
4
University of Dundee, Scotland, UK evapotranspiration (ET) and 17% (215 mm) contributes to groundwater
5
UK Centre for Ecology and Hydrology,
recharge and surface storage. Rainfall varies from 700 to 5400 mm and the
Wallingford, UK
Western Ghats (mountain pass) are the main source of freshwater genera-
Correspondence tion. The estimated ET in different catchments ranged from 500 to 900 mm
Dr Kaushal K. Garg, ICRISAT
Development Centre, International Crops
per annum. An increase in the allocation of fresh water supplied by all three
Research Institute for the Semi-Arid reservoirs (Hemavathi, Harangi and KRS) was observed in the canal com-
Tropics, Hyderabad-502 324, Telangana, mand areas, from 1450 million cubic metres (MCM) yr‾¹ in 1971–1980 to
India.
Email: [email protected]
3800 MCM yr‾¹ in 2001–2010. AWM interventions harvested 140–160 MCM
(13–20 mm) of surface runoff upstream of the upper Cauvery and reduced
Funding information
inflow into the Krishnaraja Sagar reservoir by 2–6%. The study findings are
Ministry of Earth Sciences; Natural
Environment Research Council useful for designing and planning suitable water management interventions
at basin scale.

KEYWORDS
catchment treatment, reservoir inflow, surface runoff, water balance

Résumé
Le bassin de Cauvery dans le sud de l'Inde connaît des problèmes trans-
frontaliers en raison de l'augmentation de la demande en eau. Cette étude
a analysé les composantes du bilan hydrique et l'impact des interventions
de gestion de l'eau agricole (AWM) dans le bassin versant de la Haute

* Impact des interventions de gestion de l'eau agricole sur les compromis amont-aval dans le bassin supérieur du Cauvery, dans le sud de l'Inde: une
étude de modélisation.

Irrig. and Drain. 2021;1–23. wileyonlinelibrary.com/journal/ird © 2021 John Wiley & Sons, Ltd. 1
2 WABLE ET AL.

Cauvery du bassin de la Cauvery. Les résultats ont montré que le bassin de

l'étude reçoit en moyenne 1280 mm de précipitations annuelles. Sur ce
total, 29% (370 mm) s'écoulent en aval, 54% (700 mm) contribuent à
l'évapotranspiration (ET) et 17% (215 mm) contribuent à la recharge des
eaux souterraines et au stockage en surface. Les précipitations varient de
700 à 5400 mm et les Ghâts occidentaux (col de montagne) sont la pri-
ncipale source de production d'eau douce. L'ET estimée dans différents
bassins versants variait de 500 à 900 mm par an. Une augmentation de
l'allocation d'eau douce fournie par les trois réservoirs (Hemavathi, Harangi
et KRS) a été observée dans les zones de commande du canal, de 1450 mil-
lions de mètres cubes (MCM)/an en 1971–1980 à 3800 MCM/an en 2001–
2010. Les interventions AWM ont récolté 140–160 MCM (13–20 mm) de
ruissellement de surface en amont de l'upper Cauvery et ont réduit l'afflux
dans le réservoir Krishnaraja Sagar de 2–6%. Les résultats de l'étude sont
utiles pour concevoir et planifier des interventions appropriées de gestion
de l'eau à l'échelle du bassin.

MOTS CLÉS
bilan hydrique, ruissellement de surface, traitement des bassins versants, afflux de réservoir

1 | INTRODUCTION been invested on drought mitigation measures such as

field bunds, farm ponds, check dams, terracing and
Freshwater availability is essential to ensure food secu- rejuvenating community ponds, also known as agricul-
rity for an ever-increasing population. Agriculture in tural water management (AWM) interventions (Mondal
rainfed areas is characterized by water scarcity, land et al., 2020). In situ water-harvesting interventions
degradation, low resource inputs and low productivity. (e.g. contour/graded bunds) enable improvement in soil
India's net sown area of 141 million ha of which 55% is moisture availability by enhancing the landscape's infil-
rainfed has a cropping intensity of 135%. Agricultural tration capacity, conserving moisture and controlling
productivity, generally, oscillates between 0.5 and 2.0 t soil erosion (Garg et al., 2011; Singh et al., 2014; Garg
ha‾¹ with an average of 1.0 t ha‾¹ (Rockstrom et al., 2021). Often, larger fields are divided into rela-
et al., 2009; Wani et al., 2011; Fischer, 2015; Rao tively smaller plots to reduce runoff velocity and har-
et al., 2015; Bhattacharyya et al., 2016; Fischer & vest a fraction of the runoff across the field bunds. In
Connor, 2018; Government of India [GoI], 2018; contrast, ex situ interventions harvest a fraction of sur-
Anantha et al., 2021a). Irrigated land, which constitutes face runoff that drains out from agricultural fields. Ex
45% of the total agricultural area, contributes about 55% situ interventions such as check dams and farm ponds
to the total food requirement and consumes almost 70% have a capacity varying from 100 to 10 000 m3 (Jain
of freshwater resources of the country (GoI, 2015; et al., 2007; Singh et al., 2014; Garg et al., 2020a).
Green et al., 2020). Despite concerted efforts and the investment India has
With limited scope of crop intensification in canal made in various drought mitigation measures, the
command areas, the focus has shifted towards increas- impact of AWM interventions on water balance compo-
ing groundwater (GW) recharge in dryland areas. A nents has not been fully understood (Glendenning &
number of public welfare programmes such as catch- Vervoort, 2010; Bouma et al., 2011), with some studies
ment development, the Mahatma Gandhi National focusing on one or two components of land use change
Rural Employment Guarantee Act (MGNREGA) and and crop production and others focusing on the concep-
Pradhan Mantri Krishi Sinchayee Yojana (PMKSY) have tual framework (Batchelor et al.,, 2003; Shiferaw
been initiated since 1980 as drought mitigation mea- et al., 2008).
sures (Tiwari et al., 2011; Krishnan & Studies undertaken in the water sector have
Balakrishnan, 2012; NITI Aayog, 2017, 2019; Anantha mostly focused on multi-purpose, large-scale
et al., 2021b). Since 1990, about US$14 billion have projects (major reservoirs) to address food security
WABLE ET AL. 3

(Goyal & Surampalli, 2018; Bhanja & Mukherjee, 2019); Worku et al., 2017; Mekonnen et al., 2018; Horan
mapping water use efficiency (Garg et al., 2012b); crop et al., 2021). It also allows estimation of the integrated
intensification (Jayne et al., 2004; Heller et al., 2012; impacts of changes in land use–land cover (LULC) and
Pellegrini & Fern
andez, 2018); and analysing socio- biophysical factors under different land management
economic impacts (Bhave et al., 2018; Whitehead interventions (Arnold et al., 2012; Berihun
et al., 2018); migration (Tilt et al., 2009; et al., 2020).
Deshingkar, 2012; Weinthal et al., 2015); and trans- This study aimed to analyse the impact of various
boundary issues (Sood & Mathukumalli, 2011; UNEP- AWM interventions on downstream water availability in
DHI Centre on Water and Environment, 2016). To the the upper Cauvery sub-basin of southern India. The Cau-
best of the authors' knowledge, the impact of AWM inter- very basin experiences severe water scarcity for up to
ventions on hydrological processes at catchment/basin 8 months a year, affecting over 35 million people (Ferdin
scale has not been investigated thoroughly. However, a et al., 2010; Hoekstra et al., 2012). The specific objectives
few studies have analysed their impact at micro of the study are: (i) to understand the water utilization
(<10 km2) and meso (10–100 km2) scale catchment pattern in major reservoirs of the upper Cauvery catch-
hydrology either by comparing paired catchments (Zégre ment; (ii) to analyse water balance components of the
et al., 2010; Singh et al., 2014; Sultan et al., 2018) or entire catchment; (iii) to analyse the impact of AWM
before and after catchment treatment impacts (Huang & interventions on reservoir inflow into the Krishnaraja
Zhang, 2004; Lodha & Gosain, 2007; Nyssen et al., 2009; Sagar (KRS).
Garg et al., 2011; Mekonen & Tesfahunegn, 2011). The
focus of all these studies was to quantify the impact of
catchment interventions on surface runoff, agricultural 2 | MATERIALS AND METHODS
productivity and upstream–downstream trade-offs. The
knowledge generated from micro- and meso-scale catch- 2.1 | Study area
ments was very important, but may not directly be appli-
cable to catchment- or basin-scale decision making due The Cauvery catchment is one of the largest basins in
to the difference in scale (Vinogradov et al., 2011; southern India with a catchment area of 81 155 km2.
Gentine et al., 2012). The river flows through the states of Karnataka (42.2%),
The Soil and Water Assessment Tool (SWAT) is a Tamil Nadu (54%) Kerala (3.5%) and Puducherry (0.2%)
widely used hydrological model that uses a geographic (India-WRIS WebGIS, 2014; Horan et al., 2021). The
information system (GIS) interface to capture landscape basin faces water stress and a number of socio-
variability and runs on a daily time step. SWAT has economic and political challenges. The availability of
been used to simulate water resource assessment fresh water in the basin has declined due to increasing
(Krysanova & White, 2015; Gupta et al., 2020); map population, crop intensification, industrialization and
agriculture water productivity (Garg et al., 2012b; fast urban growth over the last two decades. Competing
Thokal et al., 2015; Ahmadzadeh et al., 2016); optimize demands for water from the agricultural, domestic and
reservoir operation (Wu & Chen, 2012; Anand industrial sectors have exacerbated the situation. The
et al., 2018); study the impact of land use and manage- agriculture sector is one of the largest consumers of
ment practices (Krysanova & White, 2015; Jodar-Abellan fresh water in this basin, with agricultural land being
et al., 2019); study climate change effects (Narsimlu the major land cover type (>50%). The food security
et al., 2013; Uniyalet al., 2015; Marin et al., 2020); and and livelihood of millions of farmers mainly depend on
quantify various ecosystem services (Dile et al., 2016a, freshwater availability (both surface and groundwater
2016b; Lee et al., 2018). It has also been used to analyse resources). The catchment is characterized by large spa-
upstream–downstream water balance at meso- (Dile tial variability in terms of rainfall, land use, topography,
et al., 2016a, 2016b), catchment and basin scales (Masih soil type and various land management factors
et al., 2011). (Sreelash et al., 2020).
SWAT can capture the hydrological response to The upper Cauvery catchment was chosen to study
AWM interventions and could be customized for a the impact of AWM interventions because it is situated in
micro-scale community catchment to a large-scale the uppermost part of the basin and is relatively indepen-
catchment depending on data availability (Glavan & dent in terms of hydrological processes. The catchment
Pintar, 2012). It has been used to evaluate the impact covers 10 619 km2, 13% of the total basin. The entire
of soil conservation measures on runoff and sediment upper Cauvery catchment lies in Karnataka, covering
transport (Betrie et al., 2011; Dile et al., 2013; Dile parts of Chikkamagaluru, Kodagu, Hassan, Mandya and
et al., 2016a, 2016b; Woldesenbet et al., 2017, 2018; Mysore districts (Figure 1). Average annual rainfall in the
4 WABLE ET AL.

FIGURE 1 Upper Cauvery sub-basin, major streams, location of rain gauges, stream gauges and major reservoirs in the catchment

catchment is 1280 mm with a huge spatial variability of measured data on inflow, utilization in agriculture
600–5400 mm. The catchment includes several tributaries (canal command) and release into downstream rivers
including the Hemavathy and Laxmanthirtha, which join (section 3.2). Water balance components of the upper
the Cauvery river and flow into the KRS dam (outlet of Cauvery catchment were estimated using SWAT simu-
the study basin). The maximum storage capacity of the lation (sections 3.3 and 3.4). Using a calibrated model
KRS reservoir is 1280 million cubic metres (MCM). There set-up, the impact of AWM interventions on the KRS
are two other major reservoirs in the upper Cauvery reservoir was analysed and further projected by
basin, the Hemavathy and Harangi, with a maximum describing two future scenarios in 2030 and 2040
storage capacity of 927 and 229 MCM, respectively (section 3.5).
(Figure 1). The daily rainfall data of 23 rain gauge stations for
the period 1979–2013 (Figure 1) were collected along
with daily maximum and minimum temperature gridded
2.2 | Data collection data at the scale of 0.125 from the India Meteorological
Department (IMD). Daily relative humidity, sunshine
Figure 2 describes the methodological approach hours and wind speed for three climate stations
followed based on the study's objectives. The study (Bengaluru, Thrissur and Coimbatore) were collected for
first analysed the hydrology of the Hemavathy, the same period. The Digital Elevation Model (DEM) of
Harangi and KRS reservoirs using long-term the study area at 90 m spatial resolution was downloaded
WABLE ET AL. 5

FIGURE 2 Schema of the adopted methodology

from the CGIAR Consortium for Spatial Information 2.3 | Model description
(http://srtm.csi.cgiar.org). The land use/land cover
(LULC) map of the study area at a 1 : 250 000 scale was SWAT is a semi-process-based model that operates on a
collected from the National Remote Sensing Centre daily time step. The study catchment was divided into
(NRSC) for the year 2016. Crop statistics data were nine major land uses/land covers (Figure 3 and Table 1
obtained from the government platform (https://data. (a)). In the study area, 51% of the total geographical area
gov.in/) and web-based land use statistics (http://aps.dac. is under agriculture, 31% forest, 11% fallow/shrubland
gov.in/LUS/Index.htm). The soil map of the study area and 7% comprises settlement, water bodies and other
was acquired from the National Bureau of Soil Survey uses/covers. The land use inputs were kept the same for
and Land Use Planning (NBSS&LUP). The study also the study period (1981–2013) as there was no significant
used the soil database developed by the International change found from one to another land use between this
Crops Research Institute for the Semi-Arid Tropics period (NWDA, 2019); however, within agricultural land,
(ICRISAT) during 2005 and 2019 for previous studies in cropping intensity increased over the period. There are
Karnataka (Wani et al., 2017; Garg et al., 2020b; Anantha two major soil types: clay (6395 km2, 58%) and clay loam
et al., 2021a). (4551 km2, 42%). The entire catchment was further classi-
The daily discharge data from seven gauge locations fied into three land slopes: 0–5% covering 6415 km2
(Sakelshpur, Akkihebbal, MH Halli, Akkihebbal, Kudige, (59%), 5–10% covering 2716 km2 (25%) and greater than
Chuchunkatte and KM Vadi) were obtained from India- 10% covering 1816 km2 (16%). With all these combina-
WRIS WebGIS (http://www.india-wris.nrsc.gov.in/). tions, the entire study area was divided into 129 meso-
Information on monthly storage, inflows and outflows of scale subcatchments and 4432 hydrological response
three major reservoirs between 1970 and 2010 was units (HRUs). The daily rainfall, maximum and mini-
obtained from the Command Area Development Author- mum temperatures, relative humidity, wind speed and
ity (CADA) of Hassan, Kodagu and Mysore districts. A solar radiation between 1981 and 2013 were provided as
map of the irrigated area (command area) under these inputs to the model. Table 1(a) shows the major land use
reservoirs was obtained from the National Water Devel- classes in the study area and crop management details
opment Agency (NWDA), Bengaluru, India. Data on the provided as inputs to the model. Two major upstream
number and type of structures constructed, total treated reservoirs (Hemavathy and Harangi) were modelled by
area along with investments between 2006 and 2012 were creating the reservoir nodes at the respective sites. Their
sought from catchment development department, maximum storage capacity, water spread area and the
Bengaluru. volume required to fill the emergency spillway were
6 WABLE ET AL.

F I G U R E 3 Land use land

cover (LULC) map of the upper
Cauvery catchment (2016–17)

TABLE 1(a) Land use/land cover (LULC) statistics and crop season

Major class LULC Area (km2) Area (%) Modelled as Period

Built-up area Built up 210 2 Settlement –
Agriculture Rainy season crops 1 300 12 Sorghum, rice 15 Jun–15 Oct
Agriculture Post-rainy season crops 441 4 Sorghum 15 Nov–30 Mar
Agriculture Double/triple crops 2 443 23 Sorghum, rice 15 Jun–15 Oct, 1 Jan–31 Apr
Wasteland Current fallow 1 151 11 Rangeland Perennial
Horticulture Plantation crops 1 286 12 Coconut Perennial
Forest Forest 3 391 31 Forest Perennial
Wasteland Wasteland 119 1 Rangeland Perennial
Water Water 427 4 Water –
WABLE ET AL. 7

provided from actual records. Delineated catchments and NBSS&LUP and ICRISAT were used as direct input. Sen-
HRUs belonging to the canal command areas of the sitive parameters such as curve number, REVAP_MN,
respective reservoirs were demarcated and assigned as GWQMN and GW_DELAY that control hydrological pro-
sources of irrigation. Rice was cultivated during both cesses were used to calibrate the model.
rainy (kharif ) and post-rainy (rabi) seasons in these Reservoir nodes were created in different micro-
HRUs. An auto-irrigation rule was assigned to the model catchments to represent the AWM interventions. Based
for irrigation management. Initializing auto-irrigation on the data collected from the Department of Agricul-
enables the automatic continuation of irrigation during ture, Government of Karnataka, equivalent water-
the crop period whenever soil moisture levels are harvesting capacities were assigned both for in situ and
depleted below defined limits (Hao et al., 2015; Vories ex situ interventions. The main differences between in
et al., 2017; Chen et al., 2020). situ and ex situ interventions are the surface area, depth
Sorghum was grown during the rainy season (rice: of water harvesting and infiltration rates. Field bunds
July to November; sorghum: July to 15 November; and are common in situ interventions that harvest runoff
vegetables: July to December) and post-rainy season (rice water to a maximum height of 0.2–0.4 m, generally
and vegetables: January to April) as per LULC other than across the slope. So the water spread area is relatively
command areas. During the post-rainy period, crops were greater than in farm ponds that are excavated pits of
supported with supplemental irrigation and the ground- 2–3 m depth to harvest surface runoff (Figure 4). The
water aquifer was mapped as a source of irrigation in the water spread area to harvest 1 m3 of runoff water
drylands. Details such as date of sowing, harvesting, till- through in situ and ex situ interventions are 5–10 and
age operations and fertilizer application were provided 0.5–1.0 m2, respectively. In addition, water infiltration
based on farmers' interviews. rates of 4 mm h‾¹ in farm ponds and 12 mm h‾¹ in field
Table 1(b) shows the input values provided for the bunds were measured (based on 10 locations) at
model and their parameterization. We found that avail- Lakumanahalli micro-catchment in Chikkamagaluru
able water content (field capacity-permanent wilting district (Table 2(a)).
point) and soil depth are the most sensitive soil physical The model was run on a daily time step between 1981
parameters. Soil biophysical data retrieved from and 2013. It was calibrated by comparing simulated

TABLE 1(b) Model inputs and calibration parameters

Variable (unit) Parameter name Parameter value Source

Sand content (%) SAND 20 (10–30) NBSS&LUPa
Silt content (%) SILT 28 (20–35) NBSS&LUP
Clay content (%) CLAY 53 (35–70) NBSS&LUP
Bulk density (g cm 3) SOL_BD 1.29 (1.24–1.33) NBSS&LUP
Available water content (mm H2O (mm soil)‾¹) SOL_AWC 0.14 NBSS&LUP
Soil depth (mm) SOL_Z 750 (300–1,200) NBSS&LUP
Saturated hydraulic conductivity (mm h‾¹) SOL_K 6.6 (6.03–7.12) NBSS&LUP
Curve number CN2 82 (72–92) Calibrated
b
Groundwater revapcoeff ( ) GW_REVAP 0.02 Default
c
Threshold depth of water for revap in shallow REVAP_MN 750 Calibrated
aquifer (mm H20)
Threshold depth of water in the shallow aquifer GWQMN 1 000 Calibrated
required to return flow (mm H20)
Groundwater delay time (days) GW_DELAY 31 Calibrated
Surface runoff lag coefficient SURLAG 4 Default
Base flow alpha factor ALPHA_BF 0.375 Calibrated
a
NBSS&LUP: National Bureau of soil Survey and Land Use Planning.
b
Groundwater revapcoeff: Water may move from the shallow aquifer into the overlying unsaturated zone. As GW_REVAP approaches 0, movement of water
from the shallow aquifer to the root zone is restricted. As GW_REVAP approaches 1, the rate of transfer from the shallow aquifer to the root zone approaches
the rate of potential evapotranspiration.
c
REVAP_MN: Threshold depth of water in the shallow aquifer for ‘revap’ or percolation to the deep aquifer to occur (mm H2O).
8 WABLE ET AL.

FIGURE 4 Conceptual diagram of (a) field bund (in situ) and (b) farm pond (ex situ). Figures are not to scale

TABLE 2(a) Parameterization of in situ and ex situ AWM interventions

Parameters Parameter Ex situ interventions In situ interventions

Name of structure Farm pond Field bunds
Maximum water harvesting depth (m) Depth of water (h) 2.0 0.3
Cross section (m2). Refer to Figure 4 AREA 14 2.25a
Hydraulic conductivity of the reservoir bottom RES_K 4 12
(mm h 1)
Water harvesting capacity (m3) VOL 100 m3 per farm pond 90 m3 per landholdingb

Land slope = 2%.

a

Field bunds of 0.4 ha field (1 acre landholding) = 40 m.

b

surface runoff with observed flow data at seven gauging Scenario generation. Four land management scenarios
sites and inflows measured at three reservoir locations. were developed to analyse the impact of AWM interven-
The model's performance was evaluated using three tions on inflows into the KRS reservoir:
statistical indicators: root mean square error (RMSE),
Nash–Sutcliffe efficiency (NSE), and coefficient of deter- • No intervention scenario. This scenario represents the
mination (R2). A low RMSE value indicates better model control condition. All the reservoir nodes are removed
performance. The NSE values ranged from ∞ to 1, with from the model set-up (those that captured in situ and
values less than or very close to 0 indicating ‘unaccept- ex situ interventions). This scenario does not exclude
able’ or ‘poor’ model performance and values equal to the ancient tank system and Hemavathy and Harangi
1 indicating ‘perfect performance’. R2 ranged from 0 to reservoirs as these are integral parts of the catchment;
1, with a value of 0 indicating no correlation between • Current stage (2020). This is the current SWAT set-up
simulated and observed values. calibrated with existing rainwater-harvesting
WABLE ET AL. 9

interventions. Investments in in situ and ex situ inter- and Channarayapatna in Hassan District (station no. 8)
ventions were found to be in the ratio of 70 : 30 and received the lowest (720 mm). However, there was huge
current AWM density (intervention retention capacity a variation in the temporal scale, as shown in Figure 5
per ha) implemented in drylands was 20–25 m3 ha‾¹. (b). Overall, the average annual rainfall of the study area
• Future scenario 2030. Current structure density in the was 1280 mm.
study basin is 25 m3 ha‾¹. The Government of Karna-
taka is emphasizing the construction of farm ponds
and similar interventions with a minimum storage 3.2 | Decadal analysis of inflow, water
capacity of 150 m3 on smallholder farmers’ fields (less uses and downstream release in major
than 2.0 ha of farmland) under the farm pond scheme reservoirs
(Government of Karnataka [GoK], 2014). Such inter-
ventions are likely to lead to an additional 50 m3 ha‾¹ Krishnaraja Sagar, Hemavathy and Harangi reservoirs
retention in one decade, thereby increasing rainwater- located in the catchment have been functional since
harvesting capacity to 75 m3 ha‾¹ which was consid- 1934, 1979 and 1982, respectively. A storage capacity of
ered in the simulation; about 1240 MCM was created with the Hemavathy and
• Future scenario 2040. Further, it is assumed that Harangi reservoirs during 1979–1982. The measured
harvesting intensity in the study area will reach actual inflows and major outflows (canal and river
125 m3 ha‾¹ under this scenario. releases) of the three reservoirs were analysed for four
decades: 1971–1980, 1981–1990, 1991–2000 and 2001–
Model calibration. Table 2(b) shows model inputs for 2010. The average decadal inflows into the KRS reservoir
developing no-intervention, 2030 and 2040 scenarios. The fell by a third from 5500 to 3500 MCM yr‾¹ during 1981–
model was calibrated at 10 sites (7 runoff gauges and 1990 compared to 1971–1980 due to the construction of
3 reservoirs) using the periods shown in Table 3. Follow- two upstream reservoirs (Harangi and Hemavathy).
ing the successful calibration, the model was run with Inflows into the KRS reservoir during 1991–2000 and
the above mentioned scenarios. 2001–2010 were 4200 and 4000 MCM yr‾¹, respectively.
Inflows into the Harangi reservoir over the last three
decades were 900–1000 MCM yr‾¹, and in the Hemavathy
3 | R E SUL T S reservoir 2200–2500 MCM yr‾¹ (Figure 6(a)). Not much
inter-decadal variation in inflows was observed as they
3.1 | Rainfall characterization are located on the most upstream side and receive runoff
from the Western Ghats region that has least anthropo-
Variability of measured rainfall from 23 stations between genic interference.
1979 and 2013 is presented on a yearly timescale in Fig- The annual average canal releases of the KRS,
ures 5(a) and (b). Of the 23 stations, average annual rain- Hemavathy, and Harangi reservoirs for four decades were
fall at 10 of them was less than 1000 mm; at 5 stations 47, 55 and 61% of total inflow in the KRS, Harangi and
between 1000 and 2000 mm; at 7 stations between Hemavaty and are presented in Figure 6(b). The canal
2000 and 3000 mm and at 1 station more than 4000 mm. command area of the Harangi reservoir is located in the
Bhagamandala in Kodagu District (station no. 15) upper Cauvery catchment whereas 85% of the
received the highest annual average rainfall (5400 mm) Hemavathy canal command area is located in the study

TABLE 2(b) Model inputs to capture agriculture water management scenarios

No intervention Current Future Future

stage stage stage 1 stage 2

Scenario/time period Unit Before 2000 2020 In 2030 In 2040

2
Total treated area with in situ intervention km 0 3 350 3 350 3 350
2
Total treated area with ex situ intervention km 0 2 782 2 782 2 782
Model parameter (RES_VOL) under in situ MCM 0 10 31 52
intervention
Model parameter (RES_VOL) under ex situ MCM 0 4 13 22
interventiona
a
In addition to the major reservoirs (KRS, Harangi and Hemavathy).
10 WABLE ET AL.

F I G U R E 5 (a) Spatial variation and (b) temporal variation in annual rainfall at 23 rain gauge stations of the upper Cauvery sub-basin
over 35 years (1979–2013)

catchment and the rest lies outside the basin. In contrast, (of total inflow) in 1971–1980 to 47% in 2001–2010. Simi-
the canal command area of the KRS reservoir lies larly, water utilization in agriculture (released to the
completely outside the upper Cauvery catchment. An canal command area) in the Harangi reservoir increased
increasing trend towards the release of canal water from from 30% in 1981–1990 to 55% in 2001–2010, respectively
all three reservoirs has been observed. Total surface water and in the Hemavathy reservoir it increased from 26% in
utilization (canal water release) for agriculture was 1450 1981–1990 to 61% in 2001–2010.
MCM in 1971–1980, 2500 MCM in 1981–1990, 3500 Average annual reservoir releases to downstream
MCM in 1991–2000 and 3800 MCM in 2001–2010. Of the locations for all four decades are presented in Figure 6(c).
total inflow received into the KRS reservoir, water With reduced inflows and increased canal water release,
released for the canal command area increased from 27% downstream release from the KRS reservoir declined by
WABLE ET AL. 11

F I G U R E 6 Decade-wise analysis of the

measured (a) water inflows into reservoirs,
(b) release from reservoirs into canals and
(c) release from reservoirs into river of the
Harangi, Hemavathy and KRS reservoirs (label
values indicate percentage of total inflow)

over 55%, from 3600 MCM in 1971–1980 to 1950 MCM in Chuchunkatte and Akkihebbal and inflows into the
2001–2010. Similarly in both the upstream reservoirs Hemavathy and KRS reservoirs on a monthly timescale
(Harangi and Hemavathy), water release downstream between 1981 and 2013. The flow data for Sakaleshpur,
declined from 68–69% (of total inflow) in 1981–1990 to Chuchunkatte and Akkihebbal were only available for
36–43% in 2001–2010. 2002–2014, 2008–2014 and 2002–2014 respectively. In
general, the simulated flow at monitoring locations
agreed with the observed values as well as matching
3.3 | Model performance the peaks. However, at Kudige gauging station
(Figure 1, Figure 7(a)) and inflow at Hemavathy
Figure 7 presents the model's performance by compar- (Figure 1, Figure 7(f)), simulated flow was under-
ing simulated flow with observed flow data at four out estimated. Runoff at upstream locations is generated
of seven gauging stations of Kudige, Sakaleshpur, from the Western Ghats. It is possible that the data
12 WABLE ET AL.

F I G U R E 7 Monthly time series of observed (red) and simulated (blue) streamflow/reservoir inflows for (a) Kudige gauge site,
(b) Sakleshpur, (c) Chuchunkatte, (d) Akkihebbal, (e) Hemavathy and (f) KRS reservoir

from the rain gauges did not capture the entire rainfall The model's performance statistics from all the gauging
variability of the Western Ghats region. There was a stations and reservoirs are summarized in Table 3. Out
steep gradient of rainfall from 2000 to 5000 mm within of 10 sites, R2 was found to be more than 0.75 in eight
a 100 km distance which was not captured fully due to sites, and NSE was equal to or more than 0.5 at eight
limited rain gauge monitoring. The model's perfor- sites. Overall, the model was able to capture the catch-
mance in simulating inflows at a monthly scale into the ment hydrology fairly well.
KRS reservoir shows that it captured the rising limb,
peaks and recession limb of inflows quite well; how-
ever, the peaks were overpredicted for a few events. 3.4 | Water balance components
Model performance was further evaluated by estimating
RMSE (174 MCM), NSE (0.85) and R2 (0.88), indicating Rainfall is split into major water balance components:
that the model was in consonance with observed data. evapotranspiration (ET), runoff and change in water
WABLE ET AL. 13

storage. The average annual rainfall of the catchment 45% (<1000 mm), 25% (1000–2000 mm) and 30%
was 1280 mm and varied from 880 to 1880 mm between (>2000 mm) in a wet year.
1981 and 2013. Of this, 54% (700 mm) of total rainfall ET varied with rainfall distribution. ET in the Western
was split into ET (590–800 mm), 29% (370 mm) as catch- Ghats was higher than in agricultural land. A large area
ment outflow (170–630 mm) and the remainder 17% under forest in the Western Ghats and frequent rains gen-
(215 mm) as change in water storage (Figure 8(a)). In the erated significant ET (700–900 mm). In the study area,
current case, inflow to the KSR reservoir is considered as ET for about 10% of the catchment was less than 500 mm,
outflow from the catchment as KRS is located at the out- between 500 and 900 mm for 80% of the area and
let of the catchment. >900 mm for 10% of the area in a dry year. In normal and
Figure 8(b) shows the water balance components for wet years, ET for about 88–90% of the area was simulated
a wet (2007), normal (2008) and dry (2012) year. The to range from 500 to 900 mm and >900 mm for 10–12% of
annual rainfall received during wet, normal and dry years the area of the upper Cauvery. Groundwater withdrawal in
was 1690, 1400 and 1120 mm, respectively. Most of the agriculture was simulated from 40 to 66 mm (425–700
rainfall went towards ET, estimated to be 600–750 mm, MCM) with average of 50 mm (i.e. 530 MCM). Groundwa-
which is 40–60% of the total rainfall received. The surface ter withdrawal reported by the National Water Develop-
runoff generated was 715 mm (42% of rainfall) in a wet ment Agency for the upper Cauvery catchment in 2011
year, 450 mm (34%) in a normal year and 325 mm (29%) was 488 MCM (i.e. 46 mm) which is in close agreement
in a dry year. The change in groundwater recharge was with current analysis (GoI, 2019).
in the order of 130–230 MCM, of which 11–14% was gen- Change in groundwater storage was mapped on a spa-
erated by the rainfall received. A comparison of dry, nor- tial scale for selected dry, normal and wet years. Results
mal and wet years showed that the most sensitive water from the model showed a negative groundwater balance
balance component is surface runoff, followed by ground- in more than 50% of the area in a dry year, as water with-
water recharge with changing rainfall conditions from drawal in these catchments had been higher than the
year to year. recharge. A negative groundwater balance was found in
Figure 9 shows the spatial variability in major water 20% of the area in normal years and 3% in wet years.
balance components (rainfall, ET, runoff and change in Runoff, an important source of fresh water, was found to
groundwater storage) for the selected wet (2007), normal be the most sensitive water balance component with vari-
(2008) and dry (2012) years across the upper Cauvery. able rainfall. In a dry year, more than 50% of the catch-
The Western Ghats received the highest rainfall ment produced less than 100 mm of runoff, about 25%
(>3000 mm), with rainfall decreasing from west to east. between 100 and 500 mm, and 25% more than 500 mm.
Spatial data showed that 40% of the catchment received This proportion changed to 25% (<100 mm), 40% (100–
less than 1000 mm rainfall, 35% between 1000 and 500 mm) and 35% (>500 mm) in normal and wet years.
2000 mm, and 25% above 2000 mm during the normal Figure 10 shows the spatial variability in simulated run-
year. The distribution changed to 40% (<1000 mm), 40% off coefficients within the catchment. The runoff coeffi-
(1000–2000 mm) and 20% (>2000 mm) in a dry year and cient in 50% of the area was <0.15, and in 40% of the area

TABLE 3 Model performance statistics to simulate monthly inflows during calibration

Data availability/ Observed average Simulated average

Reservoir/ calibration monthly flow monthly flow RMSE
gauge station periods (MCM) (MCM) (MCM) PBIAS R2 NSE RSR
Hemavathy 1981–2013 193 174 127 48.1 0.81 0.79 0.46
Harangi 1981–2013 79 40 88 81.2 0.77 0.49 0.72
KRS 1981–2013 328 265 174 47.5 0.88 0.85 0.39
Akkihebbal 2002–2013 26.0 43.3 86.8 54.6 0.79 0.72 0.59
Chuchunkatte 2008–2013 169 141 90 18.1 0.88 0.86 0.37
KM Vadi 1981–2013 27 71 73 210 0.80 0.90 1.38
Kudige 1981–2013 207 157 134 33.6 0.94 0.83 0.41
MH Halli 1981–2013 93 74 81 67.0 0.67 0.60 0.64
Sakaleshpur 2002–2013 97 83 71 14.7 0.76 0.75 0.50
Thimmanahalli 2001–2013 4.95 16.4 41.1 88.1 0.30 1.16 1.61
14 WABLE ET AL.

F I G U R E 8 (a) Major water balance component (rainfall = outflow + ET + change in storage) of the upper Cauvery catchment between
1981 and 2013. The sum of the three components is equal to the total rainfall in respective years, (b) water balance in the upper Cauvery
catchment during wet (2007), normal (2008) and dry (2012) years. (Numbers in the figure indicate the percentage rainfall in a particular
year.)

0.15–0.45 during a dry year. The runoff coefficient in 30% 3.5 | Impact of AWM interventions
of the area was <0.15 and in 60% of the area 0.15–0.45.
The remaining 10% of the area had more than 0.45 runoff Figure 11 summarizes the simulated KRS reservoir inflow
coefficient during a normal year. The runoff coefficient under the four land management scenarios. Under the
for 25% of the area was less than 0.15, for 45% of the area no-intervention scenario, the annual inflows during wet,
0.15–0.45 and for 30% of the area more than 0.45 in a normal and dry years were 7800, 4300 and 3100 MCM,
wet year. respectively. Under the 2020 scenario, inflows fell by
WABLE ET AL. 15

F I G U R E 9 Spatial
variability in different water
balance components for selected
wet (2007), normal (2008) and
dry (2012) years
16 WABLE ET AL.

F I G U R E 1 0 Spatial variability in simulated runoff coefficients in different micro-catchments of the upper Cauvery catchment for
selected wet (2007), normal (2008) and dry (2012) years

F I G U R E 1 1 Impact of
interventions on inflows into the
KRS reservoir. Label values
indicates percentage decrease
with respect to the no-
intervention scenario

2–6% due to various water-harvesting interventions. of AWM interventions in 2030 and 2040, simulated
Under the 2030 and 2040 scenarios, simulation suggested results showed 300–440 MCM yr‾¹ (28–41 mm) and 460–
that intensifying AWM interventions would reduce KRS 610 MCM yr‾¹ (43–57 mm) of water harvested in
inflow by 6–15%. Simulation suggested greater flow upstream catchments, respectively. Simulation suggested
reduction in normal and dry years (by 10–15%) compared that AWM interventions filled 8–12 times of the total
to wet years (4–6%). storage capacity created under the current scenario. The
Figure 12(a) compares the efficacy of AWM interven- number of fillings fell with increased density of interven-
tions in terms of total water harvested in upstream catch- tions as the number of fillings in future scenarios
ments during wet, normal and dry years and under three (i.e. 2030 and 2040) was simulated to be 6–11 and 5–11
different land management scenarios (2020, 2030 and times, respectively (Figure 12(b)).
2040). Under the current scenario (2020), about 140–220 Figure 13 summarizes the water harvested in
MCM yr‾¹ fresh water was harvested which is equivalent upstream catchments under dry, normal and wet years
to 13–20 mm at catchment scale. With increased intensity and also under 2020, 2030 and 2040 AWM scenarios.
WABLE ET AL. 17

F I G U R E 1 2 (a) Water
harvested by in situ and ex situ
interventions and (b) number of
fillings during wet, normal and
wet years under 2020, 2030 and
2040 land management
scenarios

Currently (in 2020), more than 80–90% of catchments in so by using secondary data and field measurements to
the uplands are harvesting less than 25 mm runoff, reduce uncertainty in the results. The density of rain
including during wet years, and less than 10% of them gauge stations is low, approximately one rain gauge for
are harvesting runoff between 25 and 100 mm. With every 460 km2. This low density, especially in the West-
increased intensity of AWM interventions, simulation ern Ghats region, may not be able to capture the rainfall's
demonstrated that about 40–50% of the catchments spatial variability adequately. Rainfall in the Western
would harvest runoff less than or equal to 25 mm, 20– Ghats varies from 1000 to 5000 mm within a 50–100 km
30% of them would harvest between 25 and 75 mm and radius. We also realized that inflow modelled at upstream
10–20% of them would harvest more than 75 mm in reservoirs was far lower than the observed data at Har-
2030. angi. Within the model's set-up, we assumed a limited
cropping system whereas in reality there is a multiple
cropping system and associated land management. The
3.6 | Uncertainties in the model number of AWM interventions was simplified by creating
a reservoir node either of in situ or ex situ type for each
Though catchment hydrology is complex to model due to catchment. This may also generate uncertainty as the
heterogeneity in the topography, soil types, rainfall, land responses of different AWM interventions depend on
use and management practices, an effort was made to do their catchment (location), type and capacity.
18 WABLE ET AL.

F I G U R E 1 3 Spatial variability of total harvested water due to various AWM interventions for 2020, 2030 and 2040 scenarios under wet
(2007), normal (2008) and dry (2012) years, respectively
WABLE ET AL. 19

4 | DISCUSSION that water harvested from AWM interventions was equiv-

alent to one or two supplemental irrigations (25–
The study showed that a major portion of fresh water in 30 mm) which could be in the form of enhanced soil
the catchment came from the Western Ghats. The runoff moisture or blue water availability depending on the
coefficient of the Western Ghats was as high as 60–70%. local situation and management. However, the resulting
Thus, rainfall of more than 3000 mm generated over gains in crop productivity and crop intensification due to
2000 mm of runoff. Results also showed that fresh water such interventions were beyond the scope of this study.
generated from drylands was comparatively low as most In this basin, there is an apparent trade-off between local
of the rainfall was in the form of ET. More than 50% of benefits and downstream water availability. Upstream
the catchment, especially in dry and normal years, gener- development brings regional equity, as the uplands
ated 100 mm runoff or less with a runoff coefficient of mostly suffer from water scarcity, poor productivity and
15–20%. land degradation, whereas a little reduction in the flow at
Under the current scenario (2020), AWM interven- the KRS reservoir could be compensated by promoting
tions implemented in the drylands as a drought mitiga- improved water management practices.
tion strategy harvested 25–30 mm of water, while the rest Previous studies in semi-arid tropical central and
was available for downstream users. However, southern India have reported altered catchment hydrol-
catchment-scale water balance showed that flow reduc- ogy due to AWM interventions, with a reduction in run-
tion in the KRS reservoir due to AWM interventions was off of 30–50% compared to no-intervention conditions.
less than 6% of the total inflow generated. Runoff gener- However, at the same time, these catchments have trans-
ated from the Western Ghats is a major contributor to formed the landscape. A few studies have reported that
the KRS reservoir (surplus from Harangi and Hemavathy check dams and various rainwater storage structures
reservoirs). However, the increased density of AWM have harvested 5–10 times the total storage capacity
interventions could be a matter of concern for command developed in the degraded landscapes of central and
area authorities as the inflow at the KRS reservoir may western India, with rainfall ranging from 600 to 900 mm
decline by 6–15% by 2040. (Singh et al., 2014; Garg et al., 2021). AWM interventions
The results showed that water allocation in canal enhanced groundwater recharge (by 30–50%), crop pro-
command areas from all three reservoirs increased at ductivity (by 50–200%), crop intensification (by 30–50%)
the rate of 60 MCM yr‾¹. Water release from the KRS and controlled soil erosion and land degradation (by 70–
reservoir declined from 3600 MCM in 1971–80 to 1950 90%) compared to the non-intervention stage (Garg
MCM in 2001–2010, indicating a 55% reduction in the et al., 2011, 2012a; Singh et al., 2014; Garg et al., 2020a;
downstream part of the upper Cauvery basin. Of this, Garg et al., 2021). Garg et al. (2011) modelled the impact
only 2–6% is due to AWM interventions and the rest of various AWM interventions on hydrological processes
due to water allocation in canal command areas. A in the Osman Sagar catchment (736 km2) of the Musi
water balance analysis showed that runoff generated basin in the semi-arid tropics of southern India. The
from dryland areas during deficit years was relatively study reported that AWM interventions in the meso-scale
poor. Even under the no-intervention scenario, the run- catchment reduced inflow into the Osman Sagar reser-
off generated was far lower than the required demand voir by 40% but improved groundwater recharge and crop
from the canal command area. AWM interventions intensification by 30% and enhanced crop yields and farm
have, however, created a little more deficit against total incomes in upstream areas. This also reduced flow inten-
freshwater demand at KRS but at the same time it might sity and sedimentation in downstream water bodies. In
be helpful for alleviating drought in the uplands. The the trade-off between upstream and downstream, there
amount of water harvested by AWM interventions in a were more upstream benefits and relatively minor nega-
dry year was comparable to that in a wet year. Since tive impacts on downstream flow.
AWM interventions harvest a little surface runoff from Rainfed areas hold great untapped potential in terms
frequent events, there was not much difference in their of addressing food security and sustainable development
efficacy between different rainfall years. These interven- goals that can be unlocked using resource conservation
tions were found to harvest runoff 8–12 times per year technologies. At the same time, irrigated agriculture has
of its storage capacity. to keep productivity levels high despite reduced resource
AWM interventions in the drylands are meant to alle- availability which, historically, used to be met from
viate crop water stress by enhancing soil moisture avail- upstream sources. With developments upstream, down-
ability, providing life-saving supplemental irrigation stream irrigated ecosystems need to enhance water use
through locally harvested runoff, enhancing groundwater efficiency by adopting conservation measures and
recharge and crop intensification. Our analysis showed demand management strategies. This study's outcomes
20 WABLE ET AL.

will help stakeholders design and prioritize development MCM (43–57 mm) of surface runoff in the uplands,
plans for better water management in the basin. respectively. This may reduce inflow into the KRS
reservoir by 15% at most.

5 | C ON C L U S I ON S These findings are useful for stakeholders such as devel-

opment agencies, water authorities, water companies,
An analysis was done of the water balance components reservoir managers, decision makers, local authorities/
of the upper catchment of Cauvery basin. A Soil and councils and farmers’ associations to help them under-
Water Assessment Tool (SWAT) was applied to investi- stand the upstream and downstream trade-offs that will
gate the basin's hydrology. The model was calibrated at enable them to take informed decisions.
10 sites (7 runoff gauges and 3 reservoirs) between 1981
and 2013. The historical changes in inflow pattern, canal A C KN O WL ED G EME N T S
releases and downstream flow in all three main reservoirs The research underlying this paper was carried out under
were analysed using observation data sets. The model the UPSCAPE project of the Newton-Bhabha programme
was further parameterized to quantify AWM interven- ‘Sustaining Water Resources for Food, Energy and Eco-
tions in the upstream areas. In addition, four land man- system Services’, funded by the UK Natural Environment
agement scenarios representing no-intervention, current Research Council (NERC-UKRI) and India's Ministry of
status (2020), 2030 and 2040 were generated. The key Earth Sciences (MoES). Thanks are due to the canal com-
findings were as follows: mand authorities for providing reservoir data. The
CGIAR Research Programme—Water, Land and Ecosys-
• there is increasing water allocation for agriculture in tem's (WLE) support for ICRISAT scientists' time for
the canal command areas from the three main reser- writing this manuscript is duly acknowledged.
voirs. Surface water utilization for agriculture which
was 1450 MCM yr‾¹ in 1971–1980 increased to 3800 DA TA AVAI LA BI LI TY S T ATE ME NT
MCM yr‾¹ in 2001–2010. The average increase in water Data available on request from the authors
allocation for agriculture is 60 MCM yr‾¹. The
increased allocation led to a 55% decline in water RE FER EN CES
released from the KRS reservoir to the downstream Ahmadzadeh, H., Morid, S., Delavar, M. & Srinivasan, R. (2016)
river, from 3600 MCM yr‾¹ in 1971–1980 to 1950 MCM Using the SWAT model to assess the impacts of changing irri-
yr‾¹ in 2001–2010; gation from surface to pressurized systems on water productiv-
• the main source of fresh water in the catchment comes ity and water saving in the Zarrineh Rud catchment.
from the Western Ghats region, which receives more Agricultural Water Management, 175, 15–28. https://doi.org/10.
1016/j.agwat.2015.10.026
than 3000 mm of rainfall. The runoff coefficient of this
Anand, J., Gosain, A.K. & Khosa, R. (2018) Optimisation of multi-
region was more than 60%. The average annual rainfall
purpose reservoir operation by coupling Soil and Water Assess-
in the basin is 1280 mm; 29% of this generated outflow ment Tool (SWAT) and genetic algorithm for optimal operating
from the catchment; 54% was in the form of ET, and policy (Case Study: Ganga River Basin). Sustainability, 10(5),
the remaining 17% contributed to change in soil water 1660. https://doi.org/10.3390/su10051660
storage; Anantha, K.H., Garg, K.K., Cameron, A.P. & Dixit, S. (2021a) Seek-
• more than 50% of the drylands portion of the catch- ing sustainable pathways for fostering agricultural transforma-
ment receives rainfall ranging from 700 to 1200 mm. tion in Peninsular India. Environmental Research Letters, 16(4),
044032. https://doi.org/10.1088/1748-9326/abed7b
Most of it is in the form of ET and less than 15–20%
Anantha, K.H., Garg, K.K., Moses, S.D., Patil, M.D.,
generates surface runoff. The change in groundwater Sawargaonkar, G.L., Kamdi, P.J. et al. (2021b) Impact of natural
storage in the drylands was mostly negative as resource management interventions on water resources and
groundwater withdrawal exceeded recharge most of environmental services in different agroecological regions of
the time; India. Groundwater for Sustainable Development, 13, 100574.
• the model suggested that various in situ and ex situ https://doi.org/10.1016/j.gsd.2021.100574
interventions harvested surface runoff of about Arnold, J.G., Moriasi, D.N., Gassman, P.W., Abbaspour, K.C.,
140–160 MCM yr‾¹ under the current implementa- White, M.J., Srinivasan, R. et al. (2012) SWAT: model use,
calibration, and validation. Transactions of the ASABE, 55,
tion status (which is equivalent to 13–15 mm). This
1491–1508.
has reduced inflow into the KRS reservoir by less Batchelor, C.H., Rama Mohan Rao, M.S. & Manohar Rao, S. (2003)
than 6%. Increasing the density of AWM interven- Watershed development: A solution to water shortages in semi-
tions in the future (i.e. 2030 and 2040) is expected to arid India or part of the problem? Land Use and Water
harvest 300–440 MCM (28–41 mm) and 460–610 Resources Research, 3, 1–10.
WABLE ET AL. 21

Berihun, M.L., Tsunekawa, A., Haregeweyn, N., Dile, Y.T., Fischer, R.A. & Connor, D.J. (2018) Issues for cropping and agricul-
Tsubo, M., Fenta, A.A. et al. (2020) Evaluating runoff and sedi- tural science in the next 20 years. Field Crops Research, 222,
ment responses to soil and water conservation practices by 121–142. https://doi.org/10.1016/j.fcr.2018.03.008
employing alternative modeling approaches. Science of the Total Garg, K.K., Anantha, K.H., Nune, R., Akuraju, V.R., Singh, P.,
Environment, 747, 141118. https://doi.org/10.1016/j.scitotenv. Gumma, M.K. et al. (2020b) Impact of land use changes and
2020.141118 management practices on groundwater resources in Kolar dis-
Betrie, G.D., Mohamed, Y.A., Van Griensven, A. & Srinivasan, R. trict, Southern India. Journal of Hydrology: Regional Studies, 31,
(2011) Sediment management modelling in the Blue Nile 100732. https://doi.org/10.1016/j.ejrh.2020.100732
Basin using SWAT model. Hydrology and Earth System Garg, K.K., Anantha, K.H., Venkataradha, A., Dixit, S., Singh, R. &
Sciences, 15(3), 807–818. https://doi.org/10.5194/hess-15-807- Ragab, R. (2021) Impact of rainwater harvesting on hydrologi-
2011 cal processes in a fragile watershed of South Asia. Groundwater,
Bhanja, S.N. & Mukherjee, A. (2019) In-situ and satellite-based esti- 59, 839–855. https://doi.org/10.1111/gwat.13099
mates of usable groundwater storage across India: Implications Garg, K.K., Bharati, L., Gaur, A., George, B., Acharya, S., Jella, K. &
for drinking water supply and food security. Advances in Water Narasimhan, B. (2012b) Spatial mapping of agricultural water
Resources, 126, 15–23. https://doi.org/10.1016/j.advwatres.2019. productivity using the SWAT model in Upper Bhima Catch-
02.001 ment, India. Irrigation and Drainage, 61(1), 60–79. https://doi.
Bhattacharyya, R., Ghosh, B.N., Dogra, P., Mishra, P.K., Santra, P., org/10.1002/ird.618
Kumar, S. et al. (2016) Soil conservation issues in India. Sus- Garg, K.K., Karlberg, L., Barron, J., Wani, S.P. & Rockstrom, J.
tainability, 8(6), 565. https://doi.org/10.3390/su8060565 (2011) Assessing impacts of agricultural water interventions in
Bhave, A.G., Conway, D., Dessai, S. & Stainforth, D.A. (2018) Water the Kothapally watershed, Southern India. Hydrological Pro-
resource planning under future climate and socioeconomic cesses, 26(3), 387–404.
uncertainty in the Cauvery River Basin in Karnataka, India. Garg, K.K., Singh, R., Anantha, K.H., Singh, A.K.,
Water Resources Research, 54(2), 708–728. https://doi.org/10. Venkataradha, A., Barron, J. et al. (2020a) Impact of agricul-
1002/2017WR020970 tural water management interventions on hydrological pro-
Bouma, J.A., Biggs, T.W. & Bouwer, L.M. (2011) The downstream cesses, crop intensification and agricultural productivity: A
externalities of harvesting rainwater in semi-arid watersheds: study in a meso-scale watershed, Bundelkhand region, Central
An Indian case study. Agricultural Water Management, 98(7), India. Journal of Hydrology, 591, 125592. https://doi.org/10.
1162–1170. https://doi.org/10.1016/j.agwat.2011.02.010 1016/j.jhydrol.2020.125592
Chen, Y., Marek, G.W., Marek, T.H., Porter, D.O., Moorhead, J.E., Garg, K.K., Wani, S.P., Barron, J., Karlberg, L. & Rockstrom, J.
Heflin, K.R. et al. (2020) Watershed scale evaluation of an (2012a) Up-scaling potential impacts on water flows from agri-
improved SWAT auto-irrigation function. Environmental cultural water interventions: opportunities and trade-offs in the
Modelling & Software, 131, 104789. https://doi.org/10.1016/j. Osman Sagar catchment, Musi sub-basin, India. Hydrological
envsoft.2020.104789 Processes, 27(26), 3905–3921. https://doi.org/10.1002/hyp.9516
Deshingkar, P. (2012) Environmental risk, resilience and migration: Gentine, P., Troy, T.J., Lintner, B.R. & Findell, K.L. (2012) Scaling
implications for natural resource management and agriculture. in surface hydrology: Progress and challenges. Journal of Con-
Environmental Research Letters, 7(1), 015603. https://doi.org/ temporary Water Research & Education, 147(1), 28–40. https://
10.1088/1748-9326/7/1/015603 doi.org/10.1111/j.1936-704X.2012.03105.x
Dile, Y., Berndtsson, R. & Setegn, S.G. (2013) Hydrological response Glavan, M. & Pintar, M. (2012) Strengths, weaknesses, opportuni-
to climate change for Gilgel Abay River, in the Lake Tana Basin ties and threats of catchment modelling with soil and water
- upper Blue Nile Basin of Ethiopia. PLoS ONE, 8(10), 12–17. assessment tool (SWAT) model. In: Nayak, P. (Ed.) Water
https://doi.org/10.1371/journal.pone.0079296 Resources Management and Modeling. Rijeka, Croatia: InTech.
Dile, Y., Srinivasan, R. & Karlberg, L. (2016a) Assessing the impli- ISBN: 978-953-51-0246-5
cations of water harvesting intensification on upstream- Glendenning, C.J. & Vervoort, R.W. (2010) Hydrological impacts of
downstream social-ecological resilience: a case study in the rainwater harvesting (RWH) in a case study catchment: The
Lake Tana Basin. Science of the Total Environment, 542, 22–35. Arvari River, Rajasthan, India. Part 1: Field-scale impacts. Agri-
https://doi.org/10.1016/j.scitotenv.2015.10.065 cultural Water Management, 98(2), 331–342. https://doi.org/10.
Dile, Y.T., Karlberg, L., Daggupati, P., Srinivasan, R., Wiberg, D. & 1016/j.agwat.2010.09.003
Rockström, J. (2016b) Assessing the implications of water Government of India. (2015) Statistical year book, India, 2015. New
harvesting intensification on upstream–downstream ecosystem Delhi, India: Ministry of statistics and program
services: A case study in the Lake Tana basin. Science of the implementation.
Total Environment, 542, 22–35. https://doi.org/10.1016/j. Government of India (GoI). (2018) Statistical abstract of India, Min-
scitotenv.2015.10.065 istry of Agriculture and Farmers Welfare. New Delhi, India:
Ferdin, M., Görlitz, S. & Schwörer, S. (2010) Water stress in the Government of India.
Cauvery basin, south India. How current water management Government of Karnataka. (2014) Krushi Bhagya scheme. Depart-
approaches and allocation conflict constrain reform. ASIEN: ment of Agriculture, Government of Karnataka, Bengaluru
German Journal for Politics, Economy and Culture, 117, 27–44. (krushi Bhagya - Department of Agriculture (KSDA) (karna-
Fischer, R.A. (2015) Definitions and determination of crop yield, taka.gov.in).
yield gaps, and of rates of change. Field Crops Research, 182, Goyal, M.K. & Surampalli, R.Y. (2018) Impact of climate change on
9–18. https://doi.org/10.1016/j.fcr.2014.12.006 water resources in India. Journal of Environmental Engineering,
22 WABLE ET AL.

144(7), 04018054. https://doi.org/10.1061/(ASCE)EE.1943-7870. Lee, S., Yeo, I.Y., Lang, M.W., Sadeghi, A.M., McCarty, G.W.,
0001394 Moglen, G.E. & Evenson, G.R. (2018) Assessing the cumulative
Green, A.S., Dixit, S., Garg, K.K., Sandya, N.R., Singh, G., Vatta, K. impacts of geographically isolated wetlands on watershed
et al. (2020) An interdisciplinary framework for using archaeol- hydrology using the SWAT model coupled with improved
ogy, history and collective action to enhance India's agricul- wetland modules. Journal of Environmental Management, 223,
tural resilience and sustainability. Environmental Research 37–48. https://doi.org/10.1016/j.jenvman.2018.06.006
Letters, 15(10), 105021. https://doi.org/10.1088/1748-9326/ Lodha, P.P. & Gosain, A.K. (2007) Externalities in watershed man-
aba780 agement. Changes in water resources systems: methodologies
Gupta, A., Himanshu, S.K., Gupta, S. & Singh, R. (2020) Evaluation to maintain water security and ensure integrated management,
of the SWAT Model for Analysing the Water Balance Compo- Proceedings of Symposium HS3006 at IUGG2007, July 2007.
nents for the Upper Sabarmati Basin. In: Advances in Water IAHS Publ. 315. Perugia, Italy.
Resources Engineering and Management. Singapore: Springer, Marin, M., Clinciu, I., Tudose, N.C., Ungurean, C., Adorjani, A.,
pp. 141–151. Mihalache, A.L. & Cacovean, H. (2020) Assessing the vulnera-
Hao, B., Xue, Q., Marek, T.H., Jessup, K.E., Hou, X., Xu, W. et al. bility of water resources in the context of climate changes in a
(2015) Soil water extraction, water use, and grain yield by small forested watershed using SWAT: A review. Environmen-
drought-tolerant maize on the Texas High Plains. Agricultural tal Research, 184, 109330. https://doi.org/10.1016/j.envres.2020.
Water Management, 155, 11–21. https://doi.org/10.1016/j. 109330
agwat.2015.03.007 Masih, I., Maskey, S., Uhlenbrook, S. & Smakhtin, V. (2011) Impact
Heller, E., Rhemtulla, J.M., Lele, S., Kalacska, M., Badiger, S., of upstream changes in rain-fed agriculture on downstream
Sengupta, R. & Ramankutty, N. (2012) Mapping crop types, irri- flow in a semi-arid basin. Agricultural Water Management,
gated areas, and cropping intensities in heterogeneous land- 100(1), 36–45. https://doi.org/10.1016/j.agwat.2011.08.013
scapes of Southern India using multi-temporal medium- Mekonen, K. & Tesfahunegn, G.B. (2011) Impact assessment of soil
resolution imagery. Photogrammetric Engineering and Remote and water conservation measures at Medego watershed in
Sensing, 78(8), 815–827. https://doi.org/10.14358/PERS.78.8.815 Tigray, northern Ethiopia. Maejo International Journal of Sci-
Hoekstra, A.Y., Mekonnen, M.M., Chapagain, A.K., Mathews, R. ence and Technology, 5(3), 312.
E. & Richter, B.D. (2012) Global monthly water scarcity: blue Mekonnen, D.F., Duan, Z., Rientjes, T. & Disse, M. (2018) Analysis
water footprints versus blue water availability. PLoS ONE, 7(2), of combined and isolated effects of land-use and land-cover
e32688. https://doi.org/10.1371/journal.pone.0032688 changes and climate change on the upper Blue Nile River
Horan, R., Gowri, R., Wable, P.S., Baron, H., Keller, V.D.J., basin's streamflow. Hydrology and Earth System Sciences,
Garg, K.K. et al. (2021) A comparative assessment of hydrologi- 22(12), 6187–6207. https://doi.org/10.5194/hess-22-6187-2018
cal models in the Upper Cauvery catchment. Water, 13(2), 151. Mondal, B., Loganandhan, N., Patil, S.L., Raizada, A., Kumar, S. &
https://doi.org/10.3390/w13020151 Bagdi, G.L. (2020) Institutional performance and participatory
Huang, M. & Zhang, L. (2004) Hydrological responses to conserva- paradigms: Comparing two groups of watersheds in semi-arid
tion practices in a catchment of the Loess Plateau, China. region of India. International Soil and Water Conservation
Hydrological Processes, 18(10), 1885–1898. https://doi.org/10. Research, 8(2), 164–172. https://doi.org/10.1016/j.iswcr.2020.
1002/hyp.1454 04.002
India-WRIS WebGIS. (2014) Cauvery Basin report. Generation of Narsimlu, B., Gosain, A.K. & Chahar, B.R. (2013) Assessment of
database and implementation of web enabled water resources future climate change impacts on water resources of Upper
information system in the country (India-WRIS WebGIS), Min- Sind River Basin, India using SWAT model. Water Resources
istry of Water Resources, Government of India. New Delhi, Management, 27(10), 3647–3662. https://doi.org/10.1007/
India. s11269-013-0371-7
Jain, S.K., Agarwal, P.K. & Singh, V.P. (2007) Hydrology and water NITI Aayog. (2017) Road map of Pradhan Mantri Krishi Sinchayee
resources of India. Dordrecht, the Netherlands 1258: Springer. Yojna. Niti Aayog, Government of India. New Delhi, India.
Jayne, T.S., Yamano, T. & Nyoro, J. (2004) Interlinked credit and NITI Aayog. (2019) Composite water management index. Niti
farm intensification: evidence from Kenya. Agricultural Eco- Aayog, Government of India. New Delhi, India.
nomics, 31(2–3), 209–218. https://doi.org/10.1111/j.1574-0862. NWDA & Government of India (2019) Water balance study of the
2004.tb00258.x Upper Cauvery Sub-basin of the Cauvery Basin. Technical
Jodar-Abellan, A., Valdes-Abellan, J., Pla, C. & Gomariz-Castillo, F. study number. 47, New Delhi, India: National Water Develop-
(2019) Impact of land use changes on flash flood prediction ment Agency (NWDA), Government of India, pp. 1:130.
using a sub-daily SWAT model in five Mediterranean ungauged Nyssen, J., Clymans, W., Poesen, J., Vandecasteele, I., De Baets, S.,
watersheds (SE Spain). Science of the Total Environment, 657, Haregeweyn, N. & Deckers, J. (2009) How soil conservation
1578–1591. https://doi.org/10.1016/j.scitotenv.2018.12.034 affects the catchment sediment budget–a comprehensive study
Krishnan, S. & Balakrishnan, A. (2012) Impact of watershed works in the north Ethiopian highlands. Earth Surface Processes and
of MGNREGS on poverty alleviation – a micro-level study. Landforms, 34(9), 1216–1233. https://doi.org/10.1002/esp.1805
Indian Streams Research Journal, 2(7), 2230–7850. Pellegrini, P. & Fern
andez, R.J. (2018) Crop intensification, land
Krysanova, V. & White, M. (2015) Advances in water resources use, and on-farm energy-use efficiency during the worldwide
assessment with SWAT - an overview. Hydrological Sciences spread of the green revolution. Proceedings of the National
Journal, 60(5), 771–783. https://doi.org/10.1080/02626667.2015. Academy of Sciences, 115(10), 2335–2340. https://doi.org/10.
1029482 1073/pnas.1717072115
WABLE ET AL. 23

Rao, C.S., Lal, R., Prasad, J., Gopinath, K.A., Singh, R., Vories, E., Rhine, M. & Straatmann, Z. (2017) Investigating irriga-
Jakkula, V.S. et al. (2015) Potential challenges of rainfed tion scheduling for rice using variable rate irrigation. Agricul-
farming in India. Advances in Agronomy, 133, 113–181. https:// tural Water Management, 179, 314–323. https://doi.org/10.
doi.org/10.1016/bs.agron.2015.05.004 1016/j.agwat.2016.05.032
Rockstrom, J., Falkenmark, M., Karlberg, L., Hoff, H., Rost, S. & Wani, S.P., Anantha, K.H. & Garg, K.K. (2017) Soil properties, crop
Gerten, D. (2009) Future water availability for global food pro- yield and economics under integrated crop management
duction: the potential of green water for increasing resilience to practices in Karnataka, Southern India. World Development, 93,
global change. Water Resources Research, 45(7), W00A12. 43–61. https://doi.org/10.1016/j.worlddev.2016.12.012
https://doi.org/10.1029/2007WR006767 Wani, S.P., Anantha, K.H., Sreedevi, T.K., Sudi, R., Singh, S.N. &
Shiferaw, B., Reddy, V.R. & Wani, S.P. (2008) Watershed externali- D'Souza, M. (2011) Assessing the environmental benefits of
ties, shifting cropping patterns and groundwater depletion in watershed development: evidence from the Indian semi-arid
Indian semi-arid villages: The effect of alternative water pricing tropics. Journal of Sustainable Watershed Science & Manage-
policies. Ecological Economics, 67(2), 327–340. https://doi.org/ ment, 1(1), 10–20. https://doi.org/10.5147/jswsm.2011.0036
10.1016/j.ecolecon.2008.05.011 Weinthal, E., Zawahri, N. & Sowers, J. (2015) Securitizing water,
Singh, R., Garg, K.K., Wani, S.P., Tewari, R.K. & Dhyani, S.K. climate, and migration in Israel, Jordan, and Syria. Interna-
(2014) Impact of water management interventions on hydrol- tional Environmental Agreements: Politics, Law and Economics,
ogy and ecosystem services in Garhkundar-Dabar watershed of 15(3), 293–307. https://doi.org/10.1007/s10784-015-9279-4
Bundelkhand region, Central India. Journal of Hydrology, 509, Whitehead, P.G., Jin, L., Macadam, I., Janes, T., Sarkar, S.,
132–149. https://doi.org/10.1016/j.jhydrol.2013.11.030 Rodda, H.J.E. et al. (2018) Modelling impacts of climate change
Sood, A. & Mathukumalli, B.K.P. (2011) Managing international and socio-economic change on the Ganga, Brahmaputra,
river basins: reviewing India–Bangladesh transboundary water Meghna, Hooghly and Mahanadi river systems in India and
issues. International Journal of River Basin Management, 9(1), Bangladesh. Science of the Total Environment, 636, 1362–1372.
43–52. https://doi.org/10.1080/15715124.2011.553832 https://doi.org/10.1016/j.scitotenv.2018.04.362
Sreelash, K., Mathew, M.M., Nisha, N., Arulbalaji, P., Bindu, A. Woldesenbet, T.A., Elagib, N.A., Ribbe, L. & Heinrich, J. (2017)
G. & Sharma, R.K. (2020) Changes in the Hydrological Charac- Hydrological responses to land use/cover changes in the source
teristics of Cauvery River draining the eastern side of southern region of the upper Blue Nile Basin, Ethiopia. Science of the
Western Ghats, India. International Journal of River Basin Total Environment, 575, 724–741. https://doi.org/10.1016/j.
Management, 18(2), 1–14. https://doi.org/10.1080/15715124. scitotenv.2016.09.124
2020.1719119 Woldesenbet, T.A., Elagib, N.A., Ribbe, L. & Heinrich, J. (2018)
Sultan, D., Tsunekawa, A., Haregeweyn, N., Adgo, E., Tsubo, M., Catchment response to climate and land use changes in the
Meshesha, D.T. et al. (2018) Impact of soil and water upper Blue Nile sub-basins, Ethiopia. Science of the Total Envi-
conservation interventions on watershed runoff response in a ronment, 644, 193–206. https://doi.org/10.1016/j.scitotenv.2018.
tropical humid highland of Ethiopia. Environmental Manage- 06.198
ment, 61(5), 860–874. https://doi.org/10.1007/s00267-018- Worku, T., Khare, D. & Tripathi, S.K. (2017) Modeling runoff–
1005-x sediment response to land use/land cover changes using
Thokal, R.T., Gorantiwar, S.D., Kothari, M., Bhakar, S.R. & integrated GIS and SWAT model in the Beressa watershed.
Nandwana, B.P. (2015) Spatial Mapping of Agricultural Water Environment and Earth Science, 76(16), 1–14. https://doi.org/
Productivity Using the SWAT Model. Journal of the Institution 10.1007/s12665-017-6883-3
of Engineers (India): Series a, 96(1), 85–98. https://doi.org/10. Wu, Y. & Chen, J. (2012) An operation-based scheme for a multi-
1007/s40030-015-0113-3 year and multipurpose reservoir to enhance macroscale hydro-
Tilt, B., Braun, Y. & He, D. (2009) Social impacts of large dam pro- logic models. Journal of Hydrometeorology, 13(1), 270–283.
jects: A comparison of international case studies and implica- https://doi.org/10.1175/JHM-D-10-05028.1
tions for best practice. Journal of Environmental Management, Zégre, N., Skaugset, A.E., Som, N.A., McDonnell, J.J. & Ganio, L.M.
90, S249–S257. https://doi.org/10.1016/j.jenvman.2008.07.030 (2010) In lieu of the paired catchment approach: Hydrologic
Tiwari, R., Somashekhar, H.I., Parama, V.R., Murthy, I.K., model change detection at the catchment scale. Water
Kumar, M.S.M., Kumar, B.K.M. et al. (2011) MGNREGA for Resources Research, 46(11), W11544. https://doi.org/10.1029/
Environmental Service Enhancement and Vulnerability Reduc- 2009WR008601
tion: Rapid Appraisal in Chitradurga District, Karnataka.
Economic & Political Weekly, 46(20), 39–47.
UNEP-DHI Centre on Water and Environment. (2016) Trans-
boundary river basins: status and trends. Nairobi, Kenya. How to cite this article: Wable, P.S., Garg, K.K.,
Uniyal, B., Jha, M.K. & Verma, A.K. (2015) Assessing climate Nune, R., Venkataradha, A., KH, A., Srinivasan, V.
change impact on water balance components of a river basin et al. (2021) Impact of agricultural water
using SWAT model. Water Resources Management, 29(13), management interventions on upstream–
4767–4785. https://doi.org/10.1007/s11269-015-1089-5
downstream trade-offs in the upper Cauvery
Vinogradov, Y.B., Semenova, O.M. & Vinogradova, T.A. (2011) An
approach to the scaling problem in hydrological modelling: the
catchment, southern India: a modelling study.
deterministic modelling hydrological system. Hydrological Pro- Irrigation and Drainage, 1–23. Available from:
cesses, 25(7), 1055–1073. https://doi.org/10.1002/hyp.7901 https://doi.org/10.1002/ird.2662