Mapping simulated Bt cotton productivity in Punjab under changing climate Scenarios using modelling and GIS tools

Abstract

Cotton is very important cash crop of Pakistan and its yield and fiber quality is directly affected by future changing climate. Two-year experiment was conducted at Faisalabad, Sahiwal and Multan during summer season of 2014 and 2015 to estimate crop yield under changing climate scenarios and develop cotton productivity maps using modeling and GIS. The treatments were comprised of two sowing dates (1st May and 1st June), three cotton cultivars (FH-114, FH-142 and MNH-886) and three nitrogen rates (150, 200 and 250 kg ha-1). Experiment was planned in randomized complete block design with split-split plot. Statistical analysis of agronomic data collected from two-year experiment confirmed that cultivars FH-142 and MNH-886 sown with 200 kg N ha-1 on 1st May performed well under Faisalabad, Sahiwal and Multan. Crop model Decision Support System for Agro-Technology Transfer (DSSAT) simulated crop phenology, seasonal crop biomass, leaf area dynamics and seed cotton yield. The observed data was used for calibration and evaluation of CROPGRO-Cotton Model. Model calibrated well with the best treatment (1st May sown crop with 200 kg N ha-1) having Root Mean Square Error (RMSE) of 0.81 day, 407.28 kg ha-1, 448.40 kg ha-1, 0.23, 0 day for anthesis days, total dry matter, seed cotton yield, leaf area index and maturity days for cultivar FH-142, respectively. Simulation results from seasonal analysis depicted that under future climate, 16%, 34% and 45% yield loss at Faisalabad; 23%, 34% and 36% yield reduction at Sahiwal and 20%, 32% and 35% decrease in yield at Multan till end of early, mid and late century, respectively. Strategy analysis showed that timely sown cotton cultivar FH-142 at Faisalabad and Sahiwal and MNH-886 at Multan in month of May with 200 kg N ha-1 can be viable option to get maximum yield. Geographic Information System (GIS) maps of cotton productivity were generated by running model in R script with two methods. Spatial analysis with Weather generator showed that cotton yield will reduce in future all over Punjab and Dera Gazi Khan, Mianwali and Khushab districts have potential of higher seed cotton yield under 2°C rise in temperature in future. GIS maps with Metamodel showed similar results along with Sahiwal, Okara and Pakpatthan as potential districts for future cotton in Punjab. Crop Model and Geospatial maps based on simulation can be helpful tools to predict crop yield under future climate to develop site-specific adaptation strategies by adjustment of sowing dates and fertilizer with better management practices for different genotypes.