Climate change impact on rainfed versus irrigated farm systems: what is the threshold?

Abstract

Water scarcity is a matter of increasing concern in agriculture. In addition, climate change is expected to worsen this situation. In response to this, farmers are expected to increase irrigated land. Researchers therefore attempt to examine the difference in impacts on irrigated versus rainfed agricultural land. This, however, is no straightforward task given the fact that farmers pursue a continuum of farming systems from purely rainfed to purely irrigated farming systems. Farmers that fall in between the two extremes (for instance 40% rainfed, 60% irrigated) are attributed to one of the two extreme categories (pure rainfed or pure irrigation) depending on a threshold set by the researcher. For instance, a threshold of 50%, means that if a farmer irrigates more than 50%, he is attributed to the purely irrigated farm system.

This paper highlights the fact that this threshold should be set with caution because the climate change impacts determined by the researcher can differ a lot depending on the threshold set. Using a comparative continental-scale cross-sectional approach, this paper determines the impact of climate change by combining climate, soil, geographic, socio-economic, and FADN farm-level data in an ordinary least squares regression model. It determines the impact separately for irrigated and rainfed system, for all thresholds ranging from 0 to 100%.

Preliminary results show that conclusions drawn from research based on different thresholds, can differ significantly. For instance, when determining the marginal effect of temperature (MEt) for both rainfed and irrigated farm systems, we found that the MEt for rainfed systems was relatively stable between 0.12 and 0.08, independent on the threshold chosen. For irrigated systems on the other hand, the impacts showed an almost linear increase from -0.05 to 0.23 depending on the threshold chosen. Once the threshold is above 50%, the MEt of rainfed farms is lower than the MEt of irrigated farms. This implies that a researcher who has set the threshold at 30%, 50% or 70%, each would have come to significantly different conclusions (at 30% rainfed farms would be better off, at 50% both systems would have a similar impact of climate change, at 70% irrigated farms would be better off).

Similar results can be found for the marginal effect of precipitation (MEp) which ranges from -0.05 to -0.02 for rainfed systems, but decreases linearly from 0.1 to -0.1 for irrigated systems. Only at a threshold of 70%, the MEp for irrigated systems becomes lower than the MEp for rainfed systems. When the NCAR PCM climate scenario would take place in 2080, the effect of climate change would be more positive for irrigated systems if a threshold of 40% or higher is taken. For the ECHOG climate scenario, only at a threshold of 80%, irrigated farms have a more positive impact of climate change than rainfed farms.

Given these significant differences, it is important to gain more insights in the threshold values of the different farm systems.