Supervisors: Prof Andy Challinor (University of Leeds), Dr Elliot Crowley (University of Edinburgh), Dr Megan McKerchar

There is an urgent need for food production systems to become more resilient to climate change, whilst also increasing productivity and reducing agricultural greenhouse gas emissions. Climate-smartness is a term that refers to simultaneously achieving these three objectives. Measuring progress towards climate-smartness is a critical challenge (Challinor et al., 2022) and forms the focus on this studentship. Remote sensing (RS), crop modelling and machine learning (ML) have significant potential to improve regional-scale measurements of climate-smartness.

The work will focus initially on crop productivity modelling in Kenya where models are proving highly promising – using the very latest developments in using ML with RS, as part of the EU Horizon 2020 project CONFER.  Work so far has begun to form a new hybrid process-based and ML crop model with three core elements: i. crop phenological stage (pre anthesis – post anthesis – maturity); ii. Radiation Use Efficiency (RUE); and iii. daily change in harvest index (dHI/dt). NDVI data from MODIS (years 2010 – 2021) has been downloaded and partially analysed. The project will start with maize in Kenya, where significant advances have already been made, leaving a clear plan for the new crop productivity model:

• Develop a parameterisation to estimate the crop rate of change of harvest index (dHI/dt) using NDVI for each pixel in Kenya.
• Calculate RUE from end-of-season yield and biomass, using the method of Droutsas et al. (2022).
• Simulate maize using remotely-sensed NDVI in Kenya as inputs, via the dHI/dt, RUE and phenological stage parameterisations.

 

The project can then turn to the second objective of climate-smartness, that of reducing greenhouse emissions. Here, a range of options will be scoped before deciding which method(s) to develop. Options include the agricultural environmental impact calculator, Cool Farm Tool, which estimates greenhouse gas emissions, water use, and biodiversity; and climate-smart indicators, which combine assessments of the three objectives of climate-smart agriculture into a single metric (Arenas-Calle 2022, 2021; see Fig. 1 below).

 

Fig. 1. Climate-Smartness Indicator (CSI) values for standard (black circles) vs climate-smart (grey circles) crop treatments, taken from Arenas-Calles et al. (2019). Values of CSI near 1 indicate high water productivity (y axis) and low greenhouse gas intensity (x axis). The arrows link paired treatments from the same studies and the numbers close to the arrows indicate the CSI difference between them. The vertical dotted line represents the mean GHGI of studies in the dataset and horizontal dotted line, the mean WP.

The CASE studentship will provide a placement and engagement with the Cool Farm Alliance. The CFA is a not-for-profit member organisation with experience modelling environmental impact, including greenhouse gas emissions on farms for over a decade. The CFA has ~135 members that use the Cool Farm Tool, including large companies like Pepsico, Unilever, Danone and McCain and is used by thousands of farmers in over 100 countries supplying global markets. Their mission is “To enable millions of growers around the world to make more informed on-farm decisions that reduce their environmental impact.”.

References

 Arenas-Calle LN, Heinemann AB, Soler Da Silva MA, Santos AB, Ramirez-Villegas J, Whitfield S, Challinor AJ. 2022. Rice management decisions using process-based models with climate-smart indicators. Frontiers in Sustainable Food Systems. 6

Arenas-Calle LN, Ramirez-Villegas J, Whitfield S, Challinor AJ. 2021. Design of a Soil-based Climate-Smartness Index (SCSI) using the trend and variability of yields and soil organic carbon. Agricultural Systems. 190

Arenas-Calle LN, Whitfield S, Challinor AJ. 2019. A Climate Smartness Index (CSI) Based on Greenhouse Gas Intensity and Water Productivity: Application to Irrigated Rice. Frontiers in Sustainable Food Systems. 3

Challinor AJ, Arenas-Calles LN, Whitfield S. 2022. Measuring the Effectiveness of Climate-Smart Practices in the Context of Food Systems: Progress and Challenges. Frontiers in Sustainable Food Systems. 6

Droutsas, I, Challinor, AJ , Deva, CR  et al. (1 more author) (2022) Integration of machine learning into process-based modelling to improve simulation of complex crop responses. in silico Plants, 4 (2). diac017. ISSN 2517-5025