How can improvements in climate models help decision makers?
We know that climate change impacts are expected to vary depending on geographic location. For instance in Kenya, where I live, the rainfall increase suggested by most climate models is anticipated to result in changes in water quality in coastal counties; while in some inland regions, cropland is expected to become increasingly suitable for cultivation. However, at the local level, farmers report that they continue to rely on indigenous knowledge because climate information is too general to inform their decisions on matters such as when or where to irrigate. To make informed decisions, they require accurate and timely information, specific to their spatial scales of operation. Typically this has’t been available due to barriers including a lack of tailored climate information at the right spatial and time scales; and either poor or a complete lack of advisory services.
Listening to discussions during the recently concluded African Hydromet Forum, it was evident that there is a demand for customised climate services to build resilience in African communities and economies. This demand is driven by practical considerations, particularly when related to systems vulnerable to climate change. For instance, when water related infrastructure has to be refurbished and thereafter serve for many decades in an environment where there is uncertainty around the future climate. It is clear that most decisions are made without considering climate science. This has prompted a call to the climate research community to provide ‘usable science’.
Climate models, the primary tools used to estimate how climate might change in future, are able to mimic some of the atmosphere and ocean’s physical processes and the internal feedbacks within the climate system. (Read more on How to understand and interpret global climate models). They can simulate broad observed features of climate at larger spatial scales (continental and above). Climate model estimates are more accurate for some climate variables (e.g. temperature) than others. Rainfall, one of the most important variables to assess in a changing climate, is not adequately simulated in present climate models. In the context of Africa, when looking at the range of available climate models, they largely disagree in the direction of projected changes (i.e. is it getting wetter or drier?). Reliability of model estimates also reduces from larger to smaller scales as their ability to accurately represent local climate influences decreases. This is partly due to the inability of models to explicitly simulate key small-scale processes like thunderstorms and effects of orography such as mountain ranges or lakes, on local climate variations. As a result, we are unsure if there will be more or less rainfall in East Africa, or Nairobi.
As part of the Future Climate for Africa (FCFA) programme, researchers in the IMPALA consortium are pioneering development of a pan-African high resolution (4.5 km) convection permitting (CP) regional climate model, CP4-A. A CP model does a significantly better job in representing small scale processes than coarse resolution models that are often not able to capture these processes. Higher resolution (finer scale) simulations improve the model’s representation of dynamics such as the influence of mountains, and statistical properties of convection and heavy rainfall, thus creating a radical improvement in their ability to represent convection and local storms in future projections.
The CP4-A simulations are expected to fill some of the gaps in understanding the dynamics of some critical regional and local-scale processes that influence water cycles in Africa. For example, it improves the representation of the Lake Victoria circulation system which has been linked to deaths of 3,000 – 5,000 fishermen every year due to intense night-time thunderstorms. Such improvements in model performance will eventually assist in more accurate forecasts of the timing and severity of storms. Providing better data on local scale processes will empower decision-makers to help reduce risks and protect the livelihoods of the most vulnerable.
FCFA aims to improve our understanding of climate system processes so that we can better simulate them in climate models. Better representation of local-scale features in climate models is a very useful and significant step. However, no matter how detailed the results from climate models are, there will still be a range of underlying uncertainty. Recognising uncertainties and figuring out how to manage risks optimally should continue to be our approach to effective and appropriate decision making.
This blog is the first in a series looking at advances in CP4-A simulations and its applications for regional and local scale climate processes. Future blogs will also explore the implications of these improvements for improved climate services and development and disaster risk reduction applications. To stay abreast of these and other FCFA outputs, sign up to our newsletter and follow us on twitter@future_climate
G. Fosser · S. Khodayar · P. Berg (2015): Benefit of convection permitting climate model simulations in the representation of convective precipitation. Clim Dyn 44:45–60.
Liu, C., and Coauthors (2017): Continental‐scale convection – permitting modeling of the current and future climate of North America. Clim Dyn (2017) 49:71–95.
Prein, A. F., and Coauthors (2015): A review on regional convection-permitting climate modeling: Demonstrations, prospects, and challenges. Rev. Geophys., 53, 323–361, doi:10.1002/2014RG000475.
This blog was written by Zablone Owiti