Research

Invasive pest infestations have important impacts on agricultural yields and are projected to increase in the face of global warming. This study investigates the effects of encouraging individual and collective pest management practices on crop losses. We answer three questions: 1) which type of first-informed farmers are most effective at disseminating invasive species management practices, 2) how communication is affected by social and spatial dynamics, and 3) the effect of individual and collective management on mitigating crop losses. To answer such questions, we conduct an experiment in 110 communities of rural Honduras. We train three key farmers per community on how to identify, control, and collectively manage infestations. A year later, we find important effects of utilizing the disseminated pest management practices on mitigating crop losses. Farmers in areas with high infestation levels where the training took place have a significantly lower probability of suffering crop losses, relative to areas where there was no training. This result is entirely driven by the joint impact of individual and collective adoption. We are also able to show that in the same areas, communication, learning and adoption of the pest management practices significantly increased as a result of the intervention. In general, communication works best when farmers are socially connected and when there is high pest contagion probability. Our results show the potential of cooperative management for crop loss mitigation and suggest a cost-effective way for technology dissemination.

The Dry Corridor in Central America is one of the most susceptible areas to global warming worldwide because of its erratic weather patterns due to El Niño Southern Oscillation (ENSO). In this paper, we study the consequences of weather shocks that affected the area between 2012 to 2015 by relating year-to-year variations in weather with agricultural productivity and migration decisions. Our estimation relies on panel data at the household and individual level, and controls for important time-invariant determinants of agricultural productivity and migration. Extreme temperatures are found to decrease year to year bean and maize yields by 3.4% and 1%, respectively. We also find that migration decisions are influenced by weather deviations from individual expectations. Precipitation and temperature deviations from the 30-year climate normal increase migration, both domestically and internationally. We finally use a fixed-effects instrumental variables approach to evaluate jointly the effect of climate variables on agricultural yields and the effect of yield changes on future household migration decisions. Results from the instrumental variables approach suggest that extreme temperatures decrease agricultural yields and increase the probability of household migrants. Our findings reinforce the urgent need for climate resilience strategies, particularly on the most vulnerable areas worldwide.

Farmers in rural Honduras are experiencing important crop losses because of the invasive species Melanaphis Sacchari, commonly known as the Sugar Cane Aphid (SCA) in sorghum. 70% of sorghum farmers in our representative sample of the Dry Corridor of Honduras found the SCA in their sorghum crops and 49% of such sorghum farmers suffered crop losses. A map of the spatial diffusion of the infestation is currently unavailable, mainly because of the difficulties of collecting fieldwork data. We evaluate the viability of producing such mapping from remote sensing imagery. By overlapping high spatial resolution (10m) satellite imagery with plot area maps collected through fieldwork, we compare a variety of vegetation indices (VIs) with farmers' self-report on the detection of the SCA on their sorghum plots and the intensity of such infestations. We find the Normalized Difference Vegetation Index (NDVI) to be best suited to identify SCA infestations at the plot level and a significant decrease of the NDVI during the late stage of crop development. We also find that SCA infestations can be predicted with 65% accuracy overall and with 82% accuracy for the plots with most intense infestations. Our results are encouraging for the potential of high-resolution spatial imagery as a cost-effective method to map Melanaphis Sacchari infestations and its potential use of early warning systems or performance incentive mechanisms.