Main content area

Explaining rice yields and yield gaps in Central Luzon, Philippines: An application of stochastic frontier analysis and crop modelling

Silva, João Vasco, Reidsma, Pytrik, Laborte, Alice G., van Ittersum, Martin K.
European journal of agronomy 2017 v.82 pp. 223-241
Oryza, crop management, crop models, crop yield, data collection, dry season, econometric models, farmers, farming systems, households, irrigation rates, irrigation water, land productivity, mineral fertilizers, nitrogen, phosphorus, potassium, surveys, wet season, Philippines
Explaining yield gaps is crucial to understand the main technical constraints faced by farmers to increase land productivity. The objective of this study is to decompose the yield gap into efficiency, resource and technology yield gaps for irrigated lowland rice-based farming systems in Central Luzon, Philippines, and to explain those yield gaps using data related to crop management, biophysical constraints and available technologies.Stochastic frontier analysis was used to quantify and explain the efficiency and resource yield gaps and a crop growth model (ORYZA v3) was used to compute the technology yield gap. We combined these two methodologies into a theoretical framework to explain rice yield gaps in farmers’ fields included in the Central Luzon Loop Survey, an unbalanced panel dataset of about 100 households, collected every four to five years during the period 1966–2012.The mean yield gap estimated for the period 1979–2012 was 3.2tonha−1 in the wet season (WS) and 4.8tonha−1 in the dry season (DS). An average efficiency yield gap of 1.3tonha−1 was estimated and partly explained by untimely application of mineral fertilizers and biotic control factors. The mean resource yield gap was small in both seasons but somewhat larger in the DS (1.3tonha−1) than in the WS (1.0tonha−1). This can be partly explained by the greater N, P and K use in the highest yielding fields than in lowest yielding fields which was observed in the DS but not in the WS. The technology yield gap was on average less than 1.0tonha−1 during the WS prior to 2003 and ca. 1.6tonha−1 from 2003 to 2012 while in the DS it has been consistently large with a mean of 2.2tonha−1. Varietal shift and sub-optimal application of inputs (e.g. quantity of irrigation water and N) are the most plausible explanations for this yield gap during the WS and DS, respectively.We conclude that the technology yield gap explains nearly half of the difference between potential and actual yields while the efficiency and resource yield gaps explain each a quarter of that difference in the DS. As for the WS, particular attention should be given to the efficiency yield gap which, although decreasing with time, still accounted for nearly 40% of the overall yield gap.