Using Weather Data to Project Climate Migration: A collision of two complex systems


shutterstock_63256468-1.jpgEvery day at aWhere, we experience the difficulties of modeling and understanding the climate. In its own way, human migration is an equally difficult system to comprehend, and unlike in climate science, migration data is in short supply - most migrants, especially the vulnerable, are the least likely to be recorded by official statistics. This makes exploring the relationship between a changing climate and human migration a particularly challenging prospect.Previous aWhere blogs have discussed ways in which the climate is influencing change in cropping patterns.[1] More and more, these trends are being identified as a driver (or potential driver) of migration flows. The theoretical pathway makes sense: When crops no longer grow in an area, or no longer grow as plentifully as they once did, farming communities are forced to seek a livelihood elsewhere.

But migration patterns are rarely so straightforward. In recent years, estimated Mexican net migration to the United States fell significantly following the Great Recession, as labor demand in the US decreased and the economic position of Mexico improved.[2] One Princeton study investigating potential for future Mexican migration to the US noted that while immigration studies typically focus on causal drivers such as geographic proximity, relative economic opportunities, social and cultural networks, and political stability, weather factors may be playing an increasingly important role in our current climate.[3]


Caption: Map produced by by Mexico’s Consejo de Poblacion (Population Council). Mexican states are marked according to estimated intensity of migration from each state, with dark red indicating very high intensity and light red indicating very low intensity.

The above map gives an official picture of migration from Mexico to the United States in 2010.[4] This map indicates that migration from Mexican states that border the United States was of low to medium volume at that time, while states around the central highlands experienced more significant outmigration. This data could fit with a climate-related story; while the northern states are relatively dry, farming there is primarily large-scale with smaller labor requirements, while small-scale rainfed farming (more vulnerable to climate shifts) is more common in agricultural areas in the central highlands. Approximately one quarter of Mexico’s population lives in the rural areas that depend primarily on rain-fed agriculture.[5]

 aWhere’s historical climate data can help shed light on where conditions have worsened for rainfed farming in recent years. A number of potential metrics and analyses are possible, but the metric presented here represents change in average soil moisture levels by state during the major planting months of the year. This metric helps to evaluate how favorable conditions were for planting and establishment of major crops.[6] The chart below shows the states that experienced more than 15% averaged decline in available soil moisture between 2008 and 2016.


In this set of states, the climate has grown significantly drier in the crucial planting months since 2008. One state experiencing high declines by this measure is Chiapas, a state in the far south of Mexico which is a major producer of maize and which is reported to have had very low outmigration to the United States in 2010. Veracruz, Mexico’s second-largest agricultural producing state, is also experiencing high declines in soil moisture, and is reported to have experienced only medium-intensity outmigration to the United States.

This is only one potential measure of worsening conditions - prevailing weather during crop growth stages other than planting and establishment can also dramatically affect yields, and could be considered in a more detailed analysis. But academic studies and reporting provide some support for the picture revealed here. One 2009 study of Chiapans who had moved to the United States showed that climate impacts on farming at home played a part in their decision.[7] The Princeton study referenced above estimated that a 10% reduction in crop yields led to a 2% change in outmigration, and used macro-climate models to create scenarios of potential future emigration.[8]

Creating a more complete, high-resolution picture of where conditions are getting worse can be valuable in identifying where climate change adaptation efforts should ideally be targeted. While aWhere does not model long-term change in the atmosphere, the same methods used in this blog to estimate change in soil moisture conditions for Mexican states can of course be downscaled to any of the 25,000+ aWhere virtual weather stations located in Mexico. An example is shown in the map below.


Should it prove true that increases in Mexican migration result from reduced cropping potential, this would undoubtedly impact the US. But it should always be remembered that the effects would be felt to a far greater degree within Mexico. Most migrants, including those of Mexico, move relatively small distances within their own countries, meaning the primary logistical, economic, and emotional impacts of migration will always be felt by local Mexican communities.[9] Policymakers and observers in both the US and Mexico will need to plan carefully and compassionately, using all available information, to ensure that those living in affected areas can thrive no matter where they are.






[6] Soil moisture is defined as (total precipitation)/(potential evapotranspiration), averaged across all aWhere virtual weather stations that fall within the state border. A rise or fall in this index across time indicates increases or decreases in available soil moisture, respectively.





Subscribe to Email Updates