Discrete stochastic programming model implemented in MPMAS for the manuscript "Intra-seasonal crop management and the potential benefit of seasonal precipitation forecasts: Farm-level embedded risk analysis in Ethiopia"

This dataset contains the MPMAS implementation of the discrete stochastic programming problem used in the article "Intra-seasonal crop management and the potential benefit of seasonal precipitation forecasts: Farm-level embedded risk analysis in Ethiopia" by Troost, C., Kayamo, S.E., Rasche, L., Berger, T. accepted for publication in the Journal of Agricultural Economics. Abstract of the article: How large is the potential benefit for farmers from obtaining forecasts of total rainfall in the upcomingseason? This critically depends on the local farm system and the crop management actions availableto farmers to respond to the forecasts issued. Studies assessing the financial value of seasonalprecipitation forecasts have so far focused almost exclusively on start-of-season crop sowingdecisions using approaches for non-embedded risk analysis. However, crop management decisionsare spread over the entire growing cycle and considering this realistically could either increase ordiminish the financial value of seasonal forecast information from a theoretical perspective.Our simulation analysis demonstrates how this research gap can be closed: We employ a state-contingent embedded risk approach combining multi-stage discrete stochastic programming andhigh- performance computing to determine optimal crop management responses (crop choice, sowingdate, tied-ridging, relay cropping and fertilization) to seasonal precipitation forecasts throughout thefull crop growth cycle.For rainfed agriculture in the Central Rift Valley of Ethiopia, we find seasonal forecasts to increaserisk-adjusted expected cash surplus by about 2-3% at realistic forecast skill levels. For a typical risk-averse subsistence farming household (0.5 ha), we calculated an increase in long-term expected cashsurplus of 58 USD PPP at 2023 prices. Intra-seasonal crop management plays an important role inforecast response. A non-embedded risk approach simulated not more than 35% of the forecast’seconomic value simulated with the embedded risk approach.