Channing Arndt, Will Farmer, Ken Strzepek and James Thurlow (2011) Working Paper No. 2011/52, UNU-WIDER, September

High regional diversity within Tanzania needs to be accounted for in any analysis of the impact of climate change on agriculture, food security and living standards.

Concerns regarding climate change has triggered a wide range of studies on the economic impacts, more specifically the impact on agricultural outputs in the future and thereby the consequential effect on food availability. These studies are of crucial importance in the identification of necessary adaptation policies in response to climate change. However, many of such studies are focussed at the global or the regional level and seldom look at the national or subnational scenarios. Given the wide diversity in impacts that can be observed even within a country or within a climatic region, it is necessary to study the impact in a more disaggregated manner than usually done.

A second and an equally important issue in the analysis of climate change impact is to not only look at the supply-side factors like crop yields and agricultural output but simultaneously also study the demand-side variables like income. This is crucial as any meaningful analysis of future food security will have to take into account the changes in the purchasing power of the population and reallocation of resources across crops or sectors along with the more generic component of domestic food output and imports.

It is with this context in mind that Channing et.al. undertake an assessment of the future impact of climate change on food security in different regions of Tanzania under different scenarios for the period 2041-50. They use the General Circulation Model (GCM) climate projections to derive four climate scenarios for the future, namely the COOL, WET, HOT and DRY. These scenarios are reached at on the basis of differential temperature, precipitation and aridity in different regions of the country. A baseline scenario is also developed using historic climate data for 1997-2006, which represents the ‘no climate change’ scenario and is purposeful for contrasting with the various climate change scenarios considered.

Using the climate projections for these four scenarios, the authors employ a crop model CLICROP for estimating the yields of different crops under the four scenarios in various regions of Tanzania. The inclusion of crop-specific factors and the effect of water-logging in this model enable the latter to make crop-yield projections that are relatively more accurate. The study includes cereal crops like maize, sorghum and millets, which are the main sources of calorie for the Tanzanian population.

Finally, the crop production estimates are fed into a Dynamic Computable General Equilibrium (DGCE) model to arrive at the final effect of climate change on various sectors in the economy, also taking into account the impact on household income as supplies of food crops vary. This helps the study to incorporate the consequent reallocation of resources between crops and sectors and its implications for the final outcome.

The results from the crop modelling indicate the substantially variable impact of climate change on crop yield across the four scenarios and sub-national regions. For example in Arusha in the northern region of Tanzania, maize yields are expected to rise by 12.66% under the WET scenario compared to the baseline ‘ no climate change’ scenario, while under the DRY and HOT scenarios, the same is expected to decline by 23.1 % and 15.05% respectively. Similarly, within the WET scenario, while the maize yields are projected to increase by 15.28% in Manyara (northern region), the yields are expected to fall by 12.38% in Tabora (central region).

In fact, out of the 25 regions studied the maize yield increase under the WET scenario for 14 and decrease for 11. Under the other scenarios, fall in maize yields are more frequent than increases due to climate change. Given this diversity in the impact of climate change on crop yields, the ultimate impact on different sectors of the economy and different categories of households is also varied. The agriculture sector is the most effected by climate change for the obvious reasons. Compared to the baseline scenario, the average annual real GDP in agriculture is 11.5% less in DRY scenario for the entire country. Within agriculture, livestock and horticulture exhibit a larger decline in terms of GDP. Also, the impact on agriculture is transmitted to certain manufacturing sectors like food processing, where the GDP is around 7.8% less than the baseline figure.

It is notable that except for the sub-regions in the northern zone and northern coast where the GDP is higher than the baseline figure under the WET scenario, the average annual real GDP is lesser than the baseline scenario for all climate scenarios in all other regions. This also translates into a lower average annual real per capita food consumption in the different regions in Tanzania. For the country as a whole, the per capita food consumption is less than the baseline by 7.95% in the DRY scenario by 2050. A part of the 11.5% decline in agricultural production is offset by higher food imports, which grow by 37.1% during 2007-50. Overall, there is an unambiguous worsening of food absorption by the population in the country due to climate change with the lower quintiles getting most adversely affected.

What is to be kept in mind while using these regional and sectoral projections for formulating adaptive policies in the future, one need to note that the model does not account for possible changes due to climate change that can occur in the global economy. For example, as prices of all imports are held constant by the model, the latter does not reveal anything about what happens if global food output also declines during this period leading to higher process for food imports. Other factors affecting world food prices like bnio-fuels and speculation may also prevent Tanzanian food imports to grow by the large dimensions that the model predicts.