In August 2022, one of the most severe floods in the history of Pakistan was triggered due to the exceptionally high monsoon rainfall. It has affected ~ 33 million people across the country. The agricultural losses in the most productive Indus plains aggravated the risk of food insecurity in the country. As part of the loss and damage (L&D) assessment methodologies, we developed an approach for evaluating crop-specific post-disaster production losses based on multi-sensor satellite data. An integrated assessment was performed using various indicators derived from pre- and post-flood images of Sentinel-1 (flood extent mapping), Sentinel-2 (crop cover), and GPM (rainfall intensity measurements) to evaluate crop-specific losses. The results showed that 2.5 million ha (18% of Sindh’s total area) was inundated out of which 1.1 million ha was cropland. The remainder of crop damage came from the extreme rainfall downpour, flash floods and management deficiencies. Thus approximately 57% (2.8 million ha) of the cropland was affected out of the 4.9 million ha of agricultural area in Sindh. The analysis indicated expected production losses of 88% (3.1 million bales), 80% (1.8 million tons), and 61% (10.5 million tons) for cotton, rice, and sugarcane. This assessment provided useful tools to evaluate the L&D of agricultural production and to develop evidence-based policies enabling post-flood recovery, rehabilitation of people and restoration of livelihood.
The impact of the 2022 floods on Pakistan’s rural communities and agriculture has been devastating, resulting in the loss of crops, livestock, and essential infrastructure. The country is now facing an unprecedented food security crisis. Nationally, Pakistan’s Sindh Province accounts for 42% of the rice production, 23% of the cotton production, and 31% of the sugarcane production. In our report, we assess potential crop production losses for these crops at the sub-district level using a remote-sensing approach based on satellite imagery. The analyses are designed to support the Government of Pakistan's rehabilitation and compensation planning processes.
This study developed a rapid rice yield estimation workflow and customized yield prediction model by integrating remote sensing and meteorological data with machine learning (ML). Several issues need to be addressed while developing a crop yield estimation model, including data quality issues, data processing issues, selecting a suitable machine learning model that can learn from few available time-series data, and understanding the non-linear relationship between historical crop yield and remote sensing and meteorological factors. This study applied a series of data processing techniques and a customized ML model to improve the accuracy of crop yield estimation at the district level in Nepal. It was found that remote sensing-derived NDVI product alone was not sufficient for accurate estimation of crop yield. After incorporating other meteorological variables into the ML models, estimation accuracy improved dramatically. Along with NDVI, the meteorological variables of rainfall, soil moisture, and evapotranspiration also exhibited a strong association with rice yield. This study also found that stacking multiple tree-based regression models together could achieve better accuracy than benchmark linear regression or standalone ML models. Due to the unique and distinct physio-geographical setting of each district, a variation in estimation accuracy from district to district could be observed. Our data processing and ML model workflow achieved an average of 92% accuracy of yield estimation with RMSE 328.06 kg/ha and MAE 317.21 kg/ha. This methodological workflow can be replicated in other study areas and the results can help the local authorities and stakeholders understand the factors affecting crop yields as well as estimating crop yield before harvesting season to ensure food security and sustainability.
In an effort to adopt emerging technologies in food security assessment through a codevelopment approach, the Government of Nepal’s Ministry of Agriculture and Livestock Development (MoALD) and the International Centre for Integrated Mountain Development’s (ICIMOD) SERVIR-HKH Initiative undertook a pilot study in Chitwan District in 2019 to jointly develop methods for satellite remote sensing and machine learning-based in-season crop assessment. MoALD experts and relevant stakeholders thoroughly reviewed the approach before the honourable minister approved it for formal use in the national-level assessment for 2020 and onwards. For wider adoption of the advanced data science methods established in the pilot study, we customised the technology by developing a digital suite of software, including GeoFairy (a mobile app to facilitate field data collection by field extension professionals at the district level) and RiceMapEngine (a simplified platform for machine learning-based crop classification to facilitate crop area map production by MoALD’s GIS Section). In the current federal governance structure of Nepal, high-quality crop maps and yield estimates will not only bridge information needs among the federal and subnational institutions but also provide a means for consistent cross-country crop status assessments and communication.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.