Abstract

Ensuring resilient and sustainable agri-food systems requires new approaches to integrate environmental monitoring, digital innovation, and policy-relevant statistics. Agriculture worldwide is increasingly exposed to shocks such as droughts, extreme climate variability, and water scarcity, which directly threaten food security and rural livelihoods. In Indonesia, East Java—home to the country’s largest rice production with more than 9.27 million tons in 2024—illustrates both the potential and vulnerabilities of intensive farming systems. Approximately 488,000 hectares of rice fields in the province remain rainfed, and during the 2024 dry season, more than 31,000 hectares were affected by drought, with nearly 8,000 hectares experiencing complete crop failure. Similar conditions are faced by many agricultural regions globally, where irrigation planning is often underfunded, fragmented, or insufficiently supported by timely and integrated data. To address these challenges, this study develops an innovative framework that combines remote sensing, geospatial analytics, Input–Output modeling, and Generative Artificial Intelligence (AI) for evidence-based irrigation planning and agricultural policy design.
The methodology introduces a Water Deficit Risk Score derived from six satellite-based indicators—precipitation, land surface temperature, soil moisture index, evapotranspiration, Normalized Difference Vegetation Index (NDVI), and solar-induced chlorophyll fluorescence (SIF). Using an entropy gain method at 500-meter spatial resolution, the score captures spatio-temporal drought dynamics across three major rice-producing districts in East Java: Malang, Lumajang, and Jember. Results reveal clear seasonal patterns: risk levels rise significantly from the first planting subround (wet season) to the third (dry season), with hotspots concentrated in upland and rain-shadow areas. Overlaying these scores with rice field maps, water availability within a 5 km radius, and existing irrigation networks provides a composite Priority Irrigation Rehabilitation Index, enabling precise targeting of high-risk yet high-potential areas.
To connect biophysical insights with socioeconomic outcomes, the study integrates spatial results into the East Java Input–Output Matrix to simulate economic multipliers of irrigation interventions. Findings suggest that targeted rehabilitation in priority areas such as Kemiri Village (Jember) could substantially boost agricultural output, increase household incomes, and create significant new employment opportunities in rural communities. Spillover effects extend to manufacturing, trade, and service sectors, confirming irrigation as both an agricultural and rural development investment. This linkage between natural resource management and economic value chains aligns with broader global priorities on resilience, poverty reduction, and inclusive growth.
The key innovation of this research lies in AgriPolicy AI, a GIS-based interactive dashboard powered by Generative AI. The platform integrates spatial visualizations, economic simulations, and a policy co-pilot that generates automated, data-driven narratives and recommendations in natural language. This design directly addresses persistent barriers to the use of statistics for policymaking, by making complex geospatial and economic information accessible to non-technical decision-makers. For instance, policymakers can query: “What are the impacts of rehabilitating irrigation in Lumajang during an El Niño year?” and receive a tailored summary combining risk maps, estimated yield gains, and economic implications. Such tools exemplify how digital agriculture and AI can bridge the gap between data production and real-time policy application.
Beyond the Indonesian case, the framework has broader international relevance. The combination of multi-source satellite data, robust statistical integration, and AI-driven policy support can be adapted to other regions facing climate shocks, fragile irrigation systems, and rural poverty. The approach advances at least four priority areas:
Shocks, risks, and resilience – by quantifying spatial drought risks and identifying adaptive irrigation strategies;
Food security and sustainable agriculture – by linking water management to rice production stability and national food systems;
Innovative data sources and digital tools – through the integration of geospatial remote sensing, official statistics, and Generative AI;
Policy relevance and rural development – by demonstrating economic multipliers of irrigation and providing tools for evidence-based decision-making.
Importantly, the study also highlights issues of data quality, interoperability, and capacity building. Remote sensing provides timely and cost-efficient coverage, but its integration with national statistics requires harmonized methodologies, validation with ground data, and institutional training. The AgriPolicy AI platform is designed as an open and adaptable tool that can strengthen the capacity of national statistical systems to monitor, analyze, and disseminate agricultural and rural statistics in line with global standards.
In conclusion, this paper demonstrates how innovative data approaches—combining satellite remote sensing, geospatial modeling, Input–Output economic analysis, and Generative AI—can support more resilient irrigation planning, strengthen food security, and create measurable rural development benefits. By operationalizing these methods into a user-friendly digital platform, the study contributes to the vision of harnessing digitalization and modern statistics for sustainable agri-food systems. While grounded in East Java, Indonesia, the framework offers a transferable model for other regions facing similar challenges of climate-induced shocks, resource constraints, and the need for data-driven agricultural policies.
Keywords: agricultural statistics, food security, resilience, irrigation, remote sensing, digital agriculture, Generative AI, geospatial integration, policy support