Ninety Percent of Fields Unsurveyed

Rice pest and disease outbreaks — brown planthopper, rice blast, bacterial leaf blight, fall armyworm — cause 10–15% annual yield losses across Asia, with catastrophic localized losses reaching 50–100% during outbreaks. Effective management requires early detection and rapid response, but surveillance systems in major rice-growing countries cover a small fraction of cultivated area. In the Philippines, government crop protection officers cover 1 officer per 4,000–10,000 hectares of rice — meaning each officer would need to visit every field once to complete a single sweep over several months. In practice, surveillance is concentrated on government research stations and accessible roadside fields, missing the interior smallholder plots where outbreaks often begin. The surveillance gap means that outbreaks are detected only after they've become visible at landscape scale — too late for the targeted, early interventions that could prevent catastrophic losses.

Southeast Asia's 100+ million smallholder rice farmers are the backbone of regional food security. When pest surveillance fails, the default response is calendar-based preventive pesticide spraying — which is expensive, ecologically damaging, and increasingly ineffective as pests develop resistance. IRRI's integrated pest management (IPM) research has shown that farmer-led, information-based pest management can reduce pesticide use by 50–70% while maintaining or increasing yields. But IPM requires pest surveillance data that smallholder farmers currently don't have access to. The absence of surveillance data forces rational farmers into the worst management strategy.

IRRI and national agricultural extension systems have trained farmers in field-level pest scouting through Farmer Field Schools — an approach that works but doesn't scale because it requires season-long, facilitator-led training for each farmer group. Smartphone-based pest identification apps have been developed (Plantix, RiceDr) but require pest identification from photographs, which fails for below-threshold populations (you can identify a pest from a photo only after it's already abundant). Remote sensing from satellites can detect vegetation stress but can't distinguish pest damage from drought, nutrient deficiency, or disease — and by the time damage is satellite-visible, the outbreak is already severe. Pheromone traps and light traps provide early detection for specific pests but require physical collection and counting, which is labor-intensive and has the same coverage problem as extension officers.

A distributed surveillance network using low-cost, automated pest detection devices deployed across representative smallholder fields — rather than only at research stations — could provide the early-warning data needed for timely IPM decisions. IRRI's own researchers have identified the key design requirements: devices must operate autonomously (solar-powered, rain-resistant, no farmer maintenance), transmit data without farmer intervention, and detect pests at below-threshold populations (before damage occurs). The technical challenge is that the most damaging pests (brown planthopper, stem borer) are small, cryptic, and active at night — requiring sensing modalities beyond simple camera traps. Acoustic monitoring, automated pheromone trap counting, and environmental proxy sensors (temperature/humidity conditions that predict outbreak risk) are promising but unvalidated at field scale.

An engineering team could prototype a low-cost, solar-powered automated insect monitoring station that uses light attraction plus acoustic or imaging sensors to count and classify common rice pests, designed for deployment in smallholder fields at <$50 per unit. A data science team could build pest outbreak prediction models using environmental proxy data (weather station data, satellite vegetation indices) that could substitute for direct pest counts where monitoring devices aren't deployed. A design team could prototype a community-based surveillance system where 1 in 20 farmers serves as a sentinel observer with a standardized reporting protocol, designing the observation protocol and communication system to make the reports useful to neighboring farmers.