Agriculture accounts for 70% of global freshwater usage, majority of which is used for irrigation. Irrigated agriculture in the Texas High Plains (THP) region of the United States highly depends on groundwater availability in the Ogallala Aquifer. Rapidly declining groundwater levels, recurring droughts and increasing climate variability are primary concerns for crop production in the THP, and they necessitate the adoption of efficient irrigation strategies. The objective of this study is to identify efficient growth-stage-based variable deficit irrigation (GS-VDI) strategies for grain sorghum (Sorghum bicolor L.) production using a prioritization scheme that assigns different priority weights for yield, crop water productivity (CWP), and irrigation under different climate variability classes. The Decision Support System for Agrotechnology Transfer Cropping System Model (DSSAT CSM) CERES-Sorghum model was used to simulate a total of 257 GS-VDI strategies, including a control scenario with 100% evapotranspiration (ET)-replacement and 256 combinations of scenarios developed by implementing four levels of ET-replacement (30%, 50%, 70%, and 90%) in four growth stages of sorghum: (i) emergence to panicle initiation, (ii) panicle initiation to boot, (iii) boot to early grain filling, and (iv) early to late grain filling. A prioritization index was developed to analyse the outcomes of these strategies under three prioritization scenarios: (i) equal priority on yield, CWP, and irrigation (PS-1), (ii) higher priority on yield and CWP (PS-2), and (iii) higher priority on CWP and irrigation (PS-3). All three prioritization schemes identified S-61 (30-90-90-30% ET replacement levels during growth stages one through four) and S-29 (30-50-90-30) strategies as the ideal GS-VDI strategies during dry and wet years, respectively. During the normal years, S-62 (30-90-90-50) was identified as the ideal strategy under PS-1 and PS-2 schemes, whereas S-61 strategy was identified as ideal in the PS-3 scheme. These ideal strategies could save 26% to 42% of irrigation water at the expense of 10% to 20% loss in yield in different years. Results from this study would be useful for producers, farm managers, and natural resource conservationists in their efforts to adopt or promote GS-VDI to maximize cropland productivity and reduce irrigation water usage from the Ogallala Aquifer.

Agricultural production systems in the Midwestern states of the United States of America are considered to be contributing significantly to Hypoxia related problems in the Gulf of Mexico. Therefore, a field study was conducted at the Iowa State University’s Northeast Research Center, near Nashua, Iowa to evaluate the effects of tillage and liquid swine manure application on NO₃-N and PO₄-P leaching losses to subsurface drain system under corn and soybean production system. Liquid swine manure was applied in the fall after crop harvest, followed by a pass of chisel plow to incorporate manure into the surface soil within 48 hours of manure application. Six-year (2001-2006) data were collected and analyzed in a randomized complete block design to determine the effects of tillage and manure application on subsurface drain flow, NO₃-N and PO₄-P leaching losses with subsurface drain water and crop yields. Results of this study indicated that treatment effects on subsurface drainage volume were significant at p=0.05 showing spatial and temporal variability of NO₃-N concentrations in subsurface drainage water (p<0.01). The NO₃-N concentrations in subsurface drain water were found to be higher in all treatments but were two-folds greater in the plots receiving manure every year. No significant effects of treatments were observed on total PO₄-P leaching losses and PO₄-P concentrations in subsurface drain water as well as on corn and soybean grain yields. Corn grain yields showed an increase of 1.8% in manure applied corn plots. No significant difference, however, was found between soybean plots receiving manure every year and soybean plots with no manure application.

Groundwater modeling is a crucial tool for simulating groundwater level behavior under future climate change scenarios, and for studying the effects of water management strategies on sustainability of groundwater resources. In this study, two types of models, namely, a physical-based numerical model called MODFLOW, and a data-driven model called Genetic Algorithm-based Multilayer Perceptron (MLP-GA), were evaluated for the reliable predictions of groundwater levels in the semi-arid region of the Karnal district, Haryana. Seven hybrid MLP-GA models were developed with different combinations of input variables such as rainfall, crop evapotranspiration, deep percolation, and irrigation water requirement. The numerical model and hybrid MLP-GA models were calibrated and validated using groundwater-level data from the pre-monsoon period. Among the hybrid models, the model M-1 with four input variables (crop evapotranspiration, rainfall, deep percolation, and applied irrigation water) and 4-29-1 (four input nodes, 29 neurons in the hidden layer, and one output node) model architecture performed the best, but the numerical model showed superiority over the MLP-GA models. The numerical model and M-1 model were used to predict future groundwater levels under projected climate change scenario. According to the numerical model, under the RCP4.5 scenario, groundwater levels in the study area were projected to decline by 7.7 meters by the year 2039 compared to the reference year of 2015. The M-1 model predicted decline of 5.0 meter by the year 2039. The study concluded that all input variables are essential for accurately simulating groundwater levels using MLP-GA models, and that the numerical model is more reliable for assessing the impact of climate change on groundwater behavior during future periods.

Aquifers underlying the arid lands of Rajasthan are under stress due to expansion in groundwater-irrigated areas. This study investigated trends in the long-term groundwater level and explored linkages of groundwater with rainfall and irrigated area in 12 districts of the arid region of Rajasthan, India using 64-year (1957-2020) rainfall data of 62 stations and 36-year (1984-2019) groundwater-level data of 4042 sites for pre-monsoon and post-monsoon seasons. Box-whisker plots of the district-wise average annual rainfalls were drawn. Trends in groundwater levels were identified by employing Mann-Kendall test and their magnitudes were quantified using Sen’s slope estimator test. Furthermore, linkages of groundwater levels with rainfall and groundwater-irrigated areas were evaluated through correlation and linear regression analysis. Results indicated that annual rainfall increased at a rate of 18 mm year^-1 over the last 2 decades. Also, annual rainfall crossed an amount of 400 mm in 7 of recent 10 years. The declining trends (p<0.05) of groundwater levels were identified at more than 50% sites mainly located in Jalor, Jhunjhunu, Pali, Sikar, Nagaur and Jodhpur districts. In contrast, rising trends in Sriganganagar, Hanumangarh and Bikaner districts were attributed to excessive canal-irrigation and poor-quality groundwater. Declining groundwater levels were more prominent in Jalor, Jhunjhunu, Nagaur and Sikar districts in post-monsoon (>0.60 m year^-1) as compared to pre-monsoon (0.40-0.50 m year^-1) season. Groundwater levels revealed poor response to rainfall as evidenced from weak linear relationship with low correlation coefficients (r) values. On the contrary, groundwater-level revealed moderate to strong linkages with irrigated areas based on r ranging from 0.58 to 0.94 and 0.68 to 0.96 during pre-monsoon and post-monsoon season, respectively. Findings of this study suggest need for adopting good strategies like harvesting surplus rain/runoff water and utilizing the harvested water judiciously for irrigation or groundwater recharge. However, this will require developing incentive-based policies to encourage farmers and other stakeholders to adopt water harvesting and help curtail excessive groundwater extraction for irrigation. In addition, reducing subsidies on electricity and promoting less water-requiring crops and adoption of water-saving technologies can be part of the policy framework.

Increased climatic variability is impacting agriculture in different ways, including water demands and availability. Present study analyses the impact of changing future climatic conditions on Water Footprints (WF) of major vegetable crops (cabbage, tomato and potato) grown in Eastern Gangetic Plains (EGP), covering the states of West Bengal, Bihar and Jharkhand, of India. A daily soil water balance model was developed to assess the blue and green water use of three major vegetable crops in EGP. The model was employed to assess the average blue and green water use at daily time step for periods pertaining to baseline (2008-2018), early (2030-31), mid (2050-51) and late (2080-81) 21^st century, under different future climate scenarios of RCP2.6, RCP4.5, RCP6.0 and RCP8.5. The study also considered yield variations of these crops under future scenarios considering the monotonic trend models. It was observed that, under baseline scenario, the WF of cabbage, potato and tomato in the study area were 481.2, 2689.2 and 434.9 Mm³ yr^-1, respectively. It was predicted that, across all the climate change scenarios and time scales, the green and blue WF of cabbage would increase by 2.75% to 6.88%, whereas in case of potato, the increase was in the range of 9.64% to 15.37%. Across all the climate scenarios and time scales, the variations in WF of tomato were projected to be comparatively lower (-3.95% to 1.37%). Looking at blue and green components, the green WF of cabbage production is likely to decrease in the range of 2.5% to 14.22% under different time scales and climate change scenarios. The green WFs of potato and tomato were projected to increase by 56.49% to 221.62% and 31.31% to 110.14%, respectively. The blue WF of cabbage would increase by 13.34% to 25.87%, while that of tomato was projected to decrease by 9.41% to 26.13%. The blue WF of potato was projected to vary in the range of -4.91% to 7.1% in future. The study clearly highlights increase in blue WF of vegetable crops underlining increased irrigation water demands under future climatic scenarios. This also calls for improving infrastructure to achieve efficient water use and better management of available water resources.

Water scarcity has become a major issue globally due to climate change and unsustainable water usage practices. As the agriculture sector is the dominant water user, water scarcity poses a high risk for global food and water security. Therefore, it is crucial to effectively manage and judiciously allocate the resources in the agricultural sector to achieve sustainable food and water security. To this, the present study was conducted to demonstrate the impacts of two water-saving irrigation methods in paddy and non-paddy crops in the Rushikulya River Basin of Eastern India. An optimization model was developed for achieving the maximum economic profit in the agricultural sector, and then solved for four different irrigation scenarios to analyze the impacts on the profits. Scenario 1 represents the current irrigation method of the area, while Scenario 2 and Scenario 3 correspond to the application of Alternate Wetting and Drying (AWD) to paddy, and sprinkler irrigation to non-paddy crops, respectively. Scenario 4 is the combination of AWD and sprinkler irrigation in paddy and non-paddy crops, respectively. The optimization problem was solved using the particle swarm optimization (PSO) algorithm. The results indicated that the net annual profit for the present irrigation condition is Rs. 8135.3 million. The net annual profit increases by 10%, 12%, and 23% under Scenarios 2, 3, and 4, respectively, as compared to the present traditional irrigation practices in the study basin. It is concluded from the study that implementing multiple interventions is more effective for achieving higher crop area while getting higher economic profit as compared to individual intervention. Overall, the framework demonstrated in this study can assist the policymakers in enhancing farmers’ socio-economic condition while ensuring judicious land and water utilization and reducing the fallow lands during Rabi season under spatio-temporally varying water availability conditions.

Developing accurate drought prediction models for drought risk assessment and management, and comprehending their effectiveness is a challenging task. This study employed artificial neural network (ANN) and M5 model tree models to forecast two drought indices, i.e., one month timescale standardized precipitation index (SPI-1) and one month timescale standardized precipitation evapotranspiration index (SPEI-1) with one-month lead time for the middle Gujarat, India using 30-year (1986-2015) gridded dataset of rainfall and temperature. The models were developed considering one to twenty hidden neurons, and trained using Levenberg-Marquardt algorithm to minimize the prediction error. Log-sigmoidal transfer functions were applied in both the hidden and output layers. Of the total data (Jan 1986-Dec 2015), 70% data (Jan 1986-Dec 2006) were used for model training, 15% data (Jan 2007-Jun 2011) for cross-validation and remaining 15% data (Jul 2011 - Dec 2015) for model testing. Results indicated that prediction accuracy of SPI-1 was better than that of SPEI-1 at one-month lead time as revealed in terms of five performance indicators, namely, coefficient of determination (R²), Nash-Sutcliffe efficiency (NSE), root mean square error (RMSE), mean absolute error (MAE), and percentage peak deviation (P_dv). It was found that ANN model for each grid was having considerably different performance in forecasting SPI-1 and SPEI-1 with R², NSE (%), RMSE, P_dv (%), and MAE having highest values as 0.693, 46.93, 0.844, 62.65, and 0.65, respectively, for SPI-1, whereas these were 0.469, 20.06, 1.07, 90.25, 0.87, respectively, for SPEI-1. It was revealed that performance of ANN models in forecasting SPI-1 was better in comparison to SPEI-1. Amongst ANN and M5 models, ANN models performed better than the M5 model tree for most of the grids selected in this study. Different ANN models with log sigmoidal activation function in the hidden as well as in the output layer for SPI-1 drought index forecast with one-month lead time were suggested for use by scientists, irrigation planners, and policy makers.