Climate Impact on Agricultural Labour Productivity (CIALP)

Most climate research related to agriculture and food systems focuses predominantly on the impact on crop yields and agricultural production losses. Although important for food security, this research completely neglects the negative impact of rising temperatures on agricultural labour productivity and the wider economy. Only recently, several studies have started to investigate this issue and showed that the economic impact of heat stress on labour, of which the majority are agricultural workers, is larger than the climate change impact on plants, especially in vulnerable regions such as sub-Saharan Africa. A major limitation of most of these studies is that they rely on national labour statistics as the main source of input data and therefore are not able to adequately capture the local change in climate on agricultural labour performance. The aim of this project is to develop, test and implement a methodology to better understand and quantify the risks of climate change on agricultural labour productivity. The proposed methodology is comparable to the state-of-the art approaches that assess the risk of climate change on crop yield, which involves overlaying ML-based high-resolution crop-type and yield maps with spatial information on climate hotspots and extreme weather events.

Project description

In the CIALP project, we used ML techniques to downscale subnational labour statistics to create high-resolution maps that show the geospatial distribution of agricultural workers for India. We selected India as a case-study because this country is and will be experiencing extreme heat stress as illustrated by recent heat waves. The maps were combined with a spatially explicit heat stress metric (wet bulb globe temperature, WBGT) and exposure response functions to assess (a) the number of agricultural workers that are affected by heat stress and (b) the related loss in productivity, both under current and future climatic conditions.

Preliminary results of the analysis indicate that existing approaches to assess the impact of heat stress do not appropriately account for the location of farmers and the labour force participation rate. As a consequence, our estimation of hours lost due to heat stress in India is much lower, but still substantial, in comparison to the existing literature.

The aim of this follow-up project is to (a) share the research findings at a conference where similar topics are presented, which allows us to build up a network of relationships in the machine learning and climate adaptation communities and (b) prepare a scientific journal article that demonstrates the WUR experience in the field of machine learning and climate adaptation and in particular with regard to the topic of the impact of heat stress on agricultural labour productivity.