Monitoring Climate And Drought Impacts On Changes Of Cotton Yield Across Lower Indus River Basin

Over time human population has increased and the demand for raw agricultural materials also increased to fulfil human needs.Cotton supplies the primary raw material to the textile industry,for maintaining the required supply,monitoring the cotton cultivated area and yield at the regional level during the growth period is essential to obtain updated and accurate information.We used MODIS 8-day enhanced vegetation index(EVI)time series to detect cotton cultivated area and a Binomial probabilistic approach to obtain the probability distribution of cotton crop occurrence in the detected areas.We applied Gaussian kriging model inside the detected cotton cultivated areas to obtain cotton yield spatial distribution patterns through crop reporting service data.The results indicated that a strong correlation(R~2=0.84)between the MODIS-derived cotton cultivated area and statistical data was achieved for all study years 2001-2020.The results were validated with field survey information,and the total accuracy 84.6%for the cotton crop area detection and 92.1%yield estimation indicate the potential of the probabilistic approaches for crop area and yield monitoring through remote sensing data.Over the past few decades,cotton productivity has become unstable in Pakistan,and climate change is one of the main factors that impact cotton yield.Due to climate change,it becomes very important to understand the change trend and its impact on cotton yield at the regional level.Here,we investigate the relationship of cotton yield variability with the climatic variability using a 15-yr moving window.The piecewise regression was fitted to obtain the trend-shifting point of climate factors.The results show that precipitation has experienced an overall decreasing trend of–0.64 mm/yr during 1991-2020,with opposing trends of–1.39 mm/yr and 1.52 mm/yr before and after the trend-shifting point,respectively.We found that cotton yield variability increased at a rate of 0.17%/yr,and this trend was highly correlated with the variability of climate factors.In addition,the multiple regression analysis explains that climate change is a dominant factor and explained 81%of the variation in cotton production in the study area from 1991-2020,while it explained 73%of the variation in production from 1991-2002 and 84%from 2002-2020.These findings reveal that climate affect the distinct temporal and spatial pattern of changes in cotton yield variability at the tehsil level.Since drought is one of the most severe natural disasters in the world and affects millions of people every year in various ways,proper monitoring and complete assessment of drought-affected areas help us to minimize its consequences.We can obtain dynamic and real-time information on multivariate factors that facilitate effective agricultural drought monitoring by using remote sensing.The Orthogonal Cotton Drought Index(OCDI)has been proposed for monitoring drought-affected areas during the cotton growing season in the lower Indus River basin Pakistan.An orthogonal cotton drought index could effectively provide cotton drought information based on information of spatial variation in different subsurface which has been obtained by applying Empirical Orthogonal Function(EOF)decomposition analysis of the Vegetation Condition Index(VCI),Temperature Condition Index(TCI),Soil moisture condition Index(SMCI),and Precipitation Condition Index(PCI).Therefore,we monitored the cotton drought through OCDI for the cotton growing season(June–September)from 2001-2020.We evaluated the efficiency of OCDI with major crop yield statistics during the study period.The results of the study revealed a strong correlation up to(72%and 66%)of OCDI with the Percentage of Precipitation Anomaly(PPA)and Yield Loss Ratio(YLR),respectively.The correlation coefficient between the orthogonal cotton drought index and Standardized Precipitation Index(SPIs)is stronger than the univariate drought indices.Moreover,our results demonstrate the capacity of OCDI to obtain quick and comprehensive spatial information at a regional scale during the cotton growth season.In summary,detection of the cotton crop by combining the 8-day MODIS EVI time series data with the field data through a binomial probabilistic model provides more realistic information about cotton cultivated areas in multi-cropping zones at the regional level in this study.Through the analysis of the spatial distribution of climate change trends and their impacts on cotton yield,policymakers,farmers,and other stakeholders can better understand,where and how climate change is affecting cotton yield.This study can guide the scientists and stakeholders in the development of appropriate adaptation strategies,such as implementing irrigation systems,changing planting dates,proactively responding to drought conditions,reducing crop losses,and optimizing resource allocation to mitigate the impact of drought on cotton yield.