India Crop Model Highlights
El Nino correlations with yield losses

Recently we released a JBA India Crop Model which has been in development since 2017. This probabilistic model makes use of innovative techniques to produce 10,000 years of stochastic yields for all major Indian Crops. The model’s output can be used by re/insurers to help with pricing and to understand their exposure within the Pradhan Mantri Fasal Bima Yojana (PMFBY), also known as the Indian Prime Minister’s Crop Insurance Scheme. The model supports a scheme aimed at increasing the financial security of Indian farmers.

As part of the development of this model, we undertook research into climate conditions impacting the growth of Indian crops, the findings of which could help improve future forecasting efforts for the Indian crop market. A key difference in this unique research is that historic ENSO phases were compared to ‘simulated’ yields, rather than de-trended empirical data. 

The model highlighted that low historic crop yields of Groundnut (aka Peanut) and Soybean are significantly correlated with years in which ENSO is in a positive phase; a key factor in these low yields appears to be the effect of El Nino on the summer monsoon and rainfed crops.

The Background

While connections between historic crop yield data and El Nino Southern Oscillation (ENSO) are nothing new, what makes the findings particularly interesting is the simulated nature of the yields.

All of the yields being analysed were entirely simulated through the use of a Physical Crop Model – a process of essentially recreating past yield estimates through modelling a crop’s daily growth based on today’s farming practice and reanalysis weather data (something we published an article about in June’s Asia Insurance Review). This is particularly interesting for those invested in this market because it hints at a potential predictive ability in Physical Crop Models to forecast future yields based on the ENSO climate signal.

Our Research

The research focused on analysing correlations between the Sea Surface Temperature (SST) anomalies of the ENSO phases, using a 3-month average of the Oceanic Nino Index (ONI) 3.4 between 1950-2013. This is the temperature anomaly of an area in the central pacific as shown below (Figure 1: Region 3.4 ONI 3.4 region, ENSO)

Indian crops are primarily grown in the Kharif (summer) season, where they are sown in March/April and then harvested in the Autumn, so their success is heavily linked to the monsoon rains during June, July and August.

For this reason, the methodology focused on comparing the JJA (June-July-August) ONI anomalies to the simulated yields. As you can see below, for the years in which El Nino is active, yield averages for Soybean for the 63-year period of comparison (1950-2013) are lower by 14% on average across all the modelled Indian districts. A similar pattern was noted for Groundnut. It’s also worth noting that in some districts, the Decision Support System for Agrotechnology Transfer (DSSAT) model (which was used to simulate agricultural crop growth) estimated drops of up to 28% during El Nino years.

Figure 2: District Yields in ENSO years (El Nino Years are always below average)

Validation

While it can often be difficult to validate comparisons made on simulated data (rather than empirical data), these findings are supported by similar studies into global changes of food production and ENSO. 

Research on the Indian Kharif (summer) crops has shown that the production of staple crops (e.g. Rice, Wheat and Peanut) are adversely affected during warm phases of ENSO in India (Selvaraju 2002, Kumar et al 2004). Moreover, more recent studies have highlighted the specific link between ENSO warm phases and reductions of Soybean and Peanut (Iizumi et al 2014) –the same crops noted as having the strongest correlations in this research.

The latter research highlighted how the El Nino phases in India can directly relate to a reduced, and sometimes even cataclysmic failure of, the summer monsoon. For predominantly rainfed crops and those particularly reliant on heavy rainfall such as Soybean and Groundnut, the reduction can have major consequences.

For more information on our new India Crop Model, get in touch or read our India Crop Model Executive Briefing.

References

Iizumi, T., Luo, J. J., Challinor, A. J., Sakurai, G., Yokozawa, M., Sakuma, H., ... & Yamagata, T. (2014). Impacts of El Niño Southern Oscillation on the global yields of major crops. Nature communications, 5, 3712.

Krishna Kumar, K., Rupa Kumar, K., Ashrit, R. G., Deshpande, N. R., & Hansen, J. W. (2004). Climate impacts on Indian agriculture. International Journal of climatology, 24(11), 1375-1393.

Selvaraju, R. (2003). Impact of El Niño–southern oscillation on Indian foodgrain production. International Journal of Climatology, 23(2), 187-206.

News &
Insights

Blog Modelling and Uncertainty - Climate Sensitivity

JBA shares insights from its climate change experts about uncertainty and modelling future flood risk.

Learn more
Blog Modelling and Uncertainty - the extra dimension

As JBA continues to consider new data and methodologies, members of the climate change team share insight into the uncertainties present in climate models and how you might handle uncertainty in future flood risk modelling. This is the first of two blogs on this subject.

Continue reading
Blog Understanding flood exposure in a hard reinsurance market

With less reinsurance readily available and costs for reinsurance expected to remain high, the importance of understanding flood exposure is more important than ever. In this latest blog JBA's Lauren Hart explore the challenges.

Continue reading
Blog How can Africa manage climate change, air pollution, and sustainable development this century?

Dr Paul Young explains the link between climate change and air quality, and what it tells us about the opportunities and challenges for Africa this century.

Continue reading