Conclusions: So long, farewell, Auf Wiedersehen,

Agriculture plays a large role in the economy across Africa and is one of the most vulnerable sectors to climate change, research needs to be conducted into this area in order to protect this livelihood. Crop models are an essential part of this climate change research (Roudier et al., 2011; Challinor 2007). Predicting how changes to temperature and rainfall patterns will influence crop yields in Africa provides essential data for agriculture mitigation strategies.

Consistency of literature.

One of the main conclusions drawn from the literature surrounds the relative importance of alterations to temperature vs rainfall patterns. A general consensus shows temperature change (predictions of approximately 1.4-1.6 degrees warming by 2050, approximately 0.2 degrees per decade (IPCC, 2013)) dominates over rainfall, having the largest negative influence on crop yields (Lobell et al., 2008; Roudier et al., 2011; Thorton et al., 2009). A warmer and drier climate is anticipated to reduce crop yields across Africa by 10-20% (Thorton et al., 2009).

Temperature patterns dominate, if rainfall remained the same, temperature would still decrease yields in West Africa by 15% (Schlenker and Lobell, 2010). Rainfall has the potential to either exaggerate or alleviate the influence of warming. This is not enough, however to overcome the issues of temperature change. Slack (2006) states demonstrates changes to rainfall compensate a 1.5 degree C temperature change for millet crop by -59% and -26% for a decrease or increasing in rainfall, retrospectively. This was supported by Gaun et al (2015).

Please note. These are generalised conclusions for Africa, these results will value for different locations and crop types. For example trends generally show a decrease in crop yield, yet rainfall can also mitigate climate change impacts. Some will witness an increase in crop yield. Within the Ethiopian highlands, surrounding Addis Ababa, Jones and Thorton (2003) predicted a “substantial localised yield increase” of up to 100% in places. Despite this the overall conclusions of the study showed a reduction in yield (maize across Africa) and in some cases too drastically (it is predicted) to prevent mitigate changes. This highlights the fact the sensitivity of crops across Africa are not uniform this fact should be notified when implementing mitigation schemes as a single policy may not be successful for a whole region.

Final words

Policy is something which has risen throughout a lot of the literature addressed. Crop models are proving very effective within the academic world but outside much work still needs to be done in order to use modelling results to implement successful policies (Bouman et al., 1996). One factor which will improve the application of agricultural policy is to involve knowledge of local farmers. The incorporation of such ideas will help reduce the gap between the needs of farmers and the academic hypothesis Stephens and Middleton (2000).

Applications of crop models: what a load of crop!

So if you haven’t gathered by now, crop models will continue to be the focus of my blogs. After doing a fair bit of reading around the topic now, I have started to wonder if crop models can be implemented in real agricultural situations, successfully. If yes, where is this evidence for this? Within academic research crop models have many applications, Decision Support System for Agrothechnology transfer being one (Jones et al., 2003). However outside this there is little evidence of the application of crop models to actual agricultural systems. Crop models, in general, are covered extensively yet the focus of these papers tend to explore model development (ie. validation). The lack of academic literature, surrounding the application of crop models, is either because crop models are not used in decision making or their use is not recorded (as others who use crop models, for example NGOs and governmental bodies, do not publish their work) (Stephens and Middleton 2000).

There is no shortage of literature discussing the suitability of crop models outside the academic bubble (Basso et al., 2005; Mathews, 2002).  The paper ‘The school of de WIT crop growth simulation models: a pedigree and historical overview’ (Bouman et al., 1996) is just one which paints a gloomy picture of the application of crop models. The study argues there have been very few cases where models have been successfully applied (exceptions including models analyzing irrigation schemes or disease management).

Stephens and Middleton (2000) suggests many of these studies fail (failure referring to the little potential for real life application) due to an inappropriate focus on scientific interests, ignoring the factual issues at hand. This is a widely accepted reason for the failure of crop model application to the real world. Many models (used for academic research purposes) tend to focus on gaining accurate representation of the system, which varies from the interests of farmers, highlighting a gap between the academic questions and practical needs (of farmers). Routine planning is a factor which is often ignored in crop modelling. Many agricultural decisions are focused around timing of the wet season (this is one of the areas which produces the most risk in farming). Seasonal timing of irrigation of sugar cane, in South Africa, can determine the total crop yield. The issue of routine is less prominent for other stake holders such as local governments who are making strategic agricultural decisions (for example in the face of climate change).

The needs and expertise of local farmers needs to be incorporated into crop modeling studies in order to make them more applicable past the academic scene. The benefits of which will help in reducing agricultural risks and maintaining sustained crop yields.