Decision Support System for Enhancing Productivity in Irrigated Saline Environment Using Remote Sensing, Modelling and GIS

Abstract

The project was implemented between August 2009 and March 2014 in the Western Yamuna Canal (WYC) command of Haryana-one of the oldest canal systems in the country. The command is divided into 8 water service circles with gross cultivated area (GCA) of 13,534 km2 covering 5 districts in full (Karnal, Panipat, Sonipat, Rohtak and Jhajjar), 7 districts in part (Yamuna Nagar, Kurukshetra, Jind, Hisar, Bhiwani, Rewari and Gurgaon) and 2205 villages. The sub-project was aimed at application of ICT for development of database and DSS program for making informed decision for improving crop yield, income and livelihoods of small farmers in six saline environments. Therefore, the sub-project has adopted the advanced Geo-IT with holistic and participatory approaches for database generation, distributed modelling and development of DSS program for generating and implementing site-specific BMP based interventions at farmers’ fields to address the issue of head-tail productivity difference, low crop yield and income in saline environments in the WYC command.

The sub-project was unique in many ways by adopting a consortia mode for project development and implementation by utilizing institutions’ expertises, bottom-up approach to problem solving, state-of-the-art Geo-IT, distributed modelling, and stakeholders’ servicing to infuse confidence on the developed Database and DSS program through demonstrations, workshops and hands-on trainings to address the issues of head-tail productivity difference and sustainability of high yield in saline/sodic conditions under Butana and Jhajjar distributaries. A series of consultations with stakeholders from two selected distributaries was conducted by applying focused group discussions, brainstorming workshops, PRA, and field days. The output of these activities was supported by baseline surveys conducted in the selected villages in both distributaries to pave the way to implementing site specific BMP based interventions generated through DSS program. Since the sub-project was a multi-institute and multi-disciplinary addressing complex issue of head-tail difference in productivity and livelihoods in Butana and Jhajjar distributaries having diverse bio-physical and socio-economic constraints in saline environment, it was necessary to develop common strategies by all project partners. This was achieved through interactive meetings, discussions, and workshops during the initial and implementation phases of the sub-project. The achievements and outcome of the sub-project were grouped into eight major themes and are presented as follows.

Development and Online Dissemination of Irri-agro Informatics Spatial Database An Irri-agro Informatics Geodatabase on bio-physical and socio-economic resources of the WYC command was developed using ArcGIS v10 from secondary source maps and data, satellite remote sensing data and GPS field surveys and is comprised of 14 key thematic layers viz. canal network with system and inflow characteristics, rainfall pattern, groundwater quality, salt-affected soils, soil texture, cropping system, terrain, waterlogging, land use, infrastructure, geology, socio-economic data, satellite data derived current land use, and digital cadastral data. The geodatabase, updated annually for the rabi crop, soil and water salinity, canal inflow and remote sensing data from 2009-2013, has characterized the bio-physical resources at entire command, district, tehsil, distributary, village, watercourse and farm levels. The characteristics of the WYC command are of 3 levels in canal network (main canal to watercourse) with system and inflow characteristics; 4 rainfall departure classes (excess, normal, deficient and scanty) during 2006-13; 6 rainfall zones - 18.6% area (<500 mm), 51.6% (500-600 mm), 17.3% (600-700 mm), 8.8% (700-800 mm), 1.6% (800-900 mm), 2.1% (>900 mm); 5 groundwater quality classes- good (38.3% area), marginal (15.2%), saline (5.3%), sodic (4.2%) and saline-sodic (37%); 2 salt-affected soils (SAS)- saline (4.0%), and sodic (14.5%); 4 soil texture classes- sand (2.4%), loamy sand (6%), sandy loam (78.6%), and loam (13%); and 5 cropping systems (rice-wheat, bajra-wheat/mustard, sorghum-wheat, cotton-wheat, and sugarcane-wheat). The Geodatabase can be queried for single or multiple attributes/ features using criteria such as monsoonal rainfall, adequacy of canal supplies, groundwater quality, saline/sodic soils, soil texture, village and other information and the district-wise crop production constraints and their spatial extent were identified. These constraints have prevailed in the large parts of Jind, Sonipat, Rohtak and Jhajjar districts which are input to crop-water-salinity-yield response model to predict the crop yield loss in six saline environments.

The geodatabase has also delineated the area of low productivity district-wise in the WYC command adopting a GIS protocol using data of canal supply, GW quality, SAS and NDVI. About 7.24% of the WYC command was affected with low productivity (988.9 km2), mainly in Rohtak, Jind and Sonipat districts. The database originally developed in ESRI’s ArcGIS proprietary format was migrated to an open source platform (Quantum GIS v1.7.4) which has allowed free distribution of the database and GIS software to the stakeholders for querying and generating value added maps. The database was further migrated to PostGIS v2.0.3 and GeoServer v2.3.0 for online dissemination and a web map service of the database was developed for online visualization and querying of multi-thematic vector layers overlaid with Google map/earth by stakeholders for identifying resource constraints at watercourse or village level.

Upscaling of Wheat Yield to Command Scale The wheat yield data from 290 crop cutting samples collected from demonstration and monitoring fields in the WYC command using GPS handset and data from Agriculture Department, Haryana were correlated with the temporal NDVI spectral profile generated from three Resourcesat-1 LISS-3 imageries (19 Dec, 5 Feb and 11 Mar) for the rabi season 2010-11. These data were analysed using GIS in tandem with spectral vegetation, salinity and waterlogging indices (NDVI, NDSI and NDWI) and were upscaled to generate the map of wheat yield variation in the WYC Command using regression technique. The wheat yield ranged from 3.51 to 4.75 t ha-1. The yield less than 4.0 t ha-1 was assessed in 56% area of the command which lies in parts of Sonipat, Jind, Rohtak, Hisar, Jhajjar and Bhiwani districts.

AquaCrop and SWAP Models for Predicting Wheat Yield and Salt Dynamics Water driven AquaCrop model 4.0 with salinity option, and SWAP model were calibrated for grain yield, water productivity and rootzone salt dynamics for three salt tolerant (KRL-1-4, 19 and 210) and one high yielding (HD-2894) wheat varieties, and four water salinities (1.5, 4, 8 and 12 dS m-1) from the experimental data of Rabi 2009-10 and were validated from the data of Rabi 2010-11. The accuracy of model prediction was evaluated by model efficiency (ME), index of agreement (d) and coefficient of determination (R2) comparing between the observed and the model simulated results. The calibrated AquaCrop model resulted in ME, d and R2 of 0.99 each for grain yield; and 0.27, 0.98 and 0.99 for water productivity, respectively, for all wheat varieties and irrigation salinity levels whereas the calibrated SWAP model resulted in ME and d of 0.96 and 0.99 for grain yield; and 0.76 and 0.93 for root zone salinity, respectively. The ME, d and R2 for the validated AquaCrop model were 0.85, 0.96, and 0.94 for grain yield, respectively, for all varieties and salinity levels whereas the ME, d and R2 for the validated SWAP model were 0.75, 0.93 and 0.95 for grain yield; and 0.95, 0.98, and 0.96 for root zone salinity, respectively. Therefore, both the models could predict the wheat yield with the acceptable accuracy for all wheat varieties and irrigation salinity levels. However, SWAP could simulate the root zone soil salinity more accurately. Therefore, both the models were integrated to DSS program and can be applied for prediction of wheat grain yield and rootzone salt dynamics in the WYC Command wherever site specific input parameters for these models are available. However, AquaCrop can be preferred for limited availability of site specific input parameters.

Crop Water Demand Driven Canal Schedule The spatial variability of weather, soil, crop, canal network and inflow of Jhajjar distributary command was generated from the Irri-agri informatics geodatabase using GIS. These information were input to CROPWAT model 8.0 and the irrigation requirement of wheat and rice grown in the distributary command was estimated. The analysis indicated an area of 5601 hectares was under rice-wheat cropping system in the distributary. The irrigation requirement of wheat at field level by CROPWAT was 254.6 mm and that of rice was 969.6 mm with effective rainfall of 55.8 and 461.8 mm during the wheat and rice growing seasons, respectively. Further, the comparison of crop water demand and canal supply as per the roster for twelve locations in Jhajjar distributary in WYC command showed that the canal supply as per roster for 2011-12 is far less (< 48%) than the demand in majority of location points. Moreover, the protocols developed for assessment of irrigation requirement can be upscaled from distributary to branch canal command.

A Standalone DSS Program for Enhancing Productivity in Irrigated Saline Environment A standalone window based DSS program v1.1 was developed in Microsoft C# programming language on .NET framework 3.5 by integrating database, key modules, crop-water-salinity-yield module, calibrated AquaCrop and SWAP models to generate and evaluate the BMP based interventions for various resource scenarios in saline environments for enhancing productivity. The developed DSS application consists of six main modules- Crop Water Demand, Canal Supply, Groundwater, Irrigation Scheduling, Modelling, and BMPs based Strategies, and three supporting modules- Database, Farmer’s Services and Help. These main modules were validated, debugged and integrated into the main user interface. The Database module displays the eight thematic data of the Irri-agro Informatics Database for assessing the six saline scenarios/constraints. The Crop Water Demand module computes the crop ET from daily weather data for 2001-2013 using Penman-Monteith method and weekly crop coefficient. The irrigation demand at watercourse outlet is thus computed from aggregation of water demand of various crops after subtracting effective rainfall and capillary water, and adding conveyance and application losses. The Canal Supply module computes the canal supply and irrigation gap to meet full crop water demand whereas the Groundwater module computes the groundwater share with or without water quality consideration. In Irrigation Scheduling module, irrigation schedules to maximize/ optimize yield are generated for wheat and other crops from one of four options- canal supply or fresh groundwater in direct or conjunctive mode, deficit irrigation, effective conjunctive mode with poor quality waters, and both water and salinity stresses.

In Modelling, a crop-water-salinity-yield response module, and a module with AquaCrop and SWAP were integrated. A crop yield response module for six prevailing saline environments in the WYC command viz, Surface water stagnation, Waterlogging, Soil salinity, Soil sodicity, Saline/sodic water irrigation, and Deficit irrigation was developed to predict the relative crop yield loss in order to generate and recommend innovative BMPs for minimizing yield loss. This module was validated from the field demo data. The relative yield loss for five major crops (wheat, barley, mustard, pearlmillet and pigeon pea) in water stagnation and waterlogging can be predicted for different duration of water stagnation/depth of waterlogging and subsequently, BMPs are recommended for minimizing yield loss. The relative yield loss in soil salinity and sodicity can be predicted for rootzone salinity (ECe) and sodicity (ESP) values at sowing, mid and harvest time for five crops. The BMPs for four ranges of ECe (< 4, 4-8, 8-12 and >12 dS m-1) and three ranges of ESP (< 20, 20-50 and >50%) were recommended for minimizing yield loss. The Soil EC converter (EC2 to ECe), and soil ESP converter (SAR, pHs and pH2 to ESP) were developed for use of data from state departments. The gypsum requirement (GR) can be computed using Schoonover’s formula or standard GR graph.

Water quality for saline/Sodic water irrigation and its permissible range for direct or conjunctive application in different agro-climatic zones are assessed. The relative yield loss can be predicted for any water salinity/sodicity values for five crops and BMPs with direct/ conjunctive use are suggested for minimizing yield loss. In deficit irrigation, the phenological growth stages for five crops are assessed and a deficit irrigation strategy based on number of available irrigations is suggested. A module developed at WTC with yield production functions, AquaCrop and SWAP was integrated under Modelling menu and can estimate the crop yield under varying soil and water salinities, foliar potassium fertilization and salt deposition. SWAP can simulate crop productivity and rootzone salinity build-up wherever site specific input parameters are available. BMP based strategies for six saline environments with their quantitative impact, and useful information for farmers on soil and water sampling procedures and testing facilities, salt tolerant and high yielding crop varieties, Help and Hindi support are also provided for use of stakeholders.

Wheat Demonstration at Farmers’ Fields In order to develop confidence of farmers on DSS generated BMP based interventions, field demonstrations of wheat crop at 52 farmers’ fields in mid and tail reaches of Butana distributary and Jhajjar distributary in saline environments were conducted during three rabi seasons (2010-11, 2011-12 and 2012-13) in Sonipat, Rohtak and Jhajjar districts. The DSS generated BMPs- four high yielding (HD-2967, 2891, and 2894, and DBW-17) and three salt tolerant varieties, (KRL-1-4, 19, and 210), optimum irrigation scheduling, effective conjunctive use of moderate saline, SAR saline and high RSC sodic groundwater, zero tillage, and laser land leveling were evaluated for enhancing crop yield. The wheat yield increased ranging from 17 to 33% in saline environment and improved the income of small farmers by Rs. 13,490-25,700 per hectare. The field demonstrations have infused confidence in stakeholders on DSS generated interventions and these results have also validated the DSS modules.

Transfer of Database, DSS Program and Knowledge to Stakeholders Since stakeholder’s servicing was the important activity of the project with 13% budget allocation, 121 district officers/engineers from CADA, Agriculture and, Irrigation Departments, KVKs and Regional Research Stations (CCS Haryana Agricultural University, Hisar) and NGOs from 12 districts within the WYC Command were imparted skill and knowledge on Database, DSS program and their application through hands-on trainings and workshops for generating BMPs for enhancing productivity in six saline environments. Similarly, 1194 members from canal water users’ associations and farmers from Karnal, Panipat, Sonipat, Jind, Rohtak, Jhajjar, Rewari and Bhiwani districts were imparted knowledge on DSS generated BMP based interventions for growing bumper crop yield under six prevailing saline environments.

Upscaling of DSS Program A feasibility assessment for upscaling of DSS program v1.1 was conducted in 7 KVKs at Panipat, Sonipat, Rohtak, Jhajjar, Rewari, Jind, and Kaithal in terms of availability of computer hardware and software, internet connectivity, manpower, and power supply with generator backup. The problems encountered were non-availability of suitable manpower at Rohtak, and irregular power supply during working hours at Panipat, Sonipat, and Jind due to rural power supply connection. The DSS program was tested at the existing computers at 3 KVKs (Panipat, Rohtak and Rewari) and the backstopping services for deployment are also being provided. Six Divisional offices of CADA at Karnal, Sonipat, Panipat, Rohtak, Jind and Kaithal were also assessed and have met all the requirements for DSS deployment. The upscaling and trainings are being continued from Institute fund and fund from CADA. The stakeholders have shown keenness on use of database and DSS program for solving their field problems.