Modelling Climate Variability and Climate Change and Their Associated Effects on Highland Cooking Banana Production in Uganda

Abstract

Climate extremes associated with climate variability and change are on the rise both in time and space. These extremes have far reaching impacts on socio-economic sectors particularly for countries like Uganda that rely heavily on rain-fed agriculture. The highland cooking banana (Musa genome group AAA-EA) is a major food crop in Uganda. Its continuous cycles of harvests makes it an important crop for enhancing food security and farmers' incomes. Studies have, however, observed continuous decline in banana productivity due to biological and environmental factors including climate extremes. This study is aimed at investigating the extent of climate variability and climate change and their associated effects on banana production over Uganda. The study used historical observed climate data (1931 to 2013), banana yields (1971 to 2009) and model simulated climate data (1991 to 2100). Climate data analysed consisted mainly of rainfall and temperature. The Providing Regional Climates for Impacts Studies (PRECIS) Regional Climate Model (RCM) was used to simulate high resolution climate projections based on the Special Report on Emission Scenarios (SRES) A1B and A2 scenarios. The study also analysed climate projection data based on the full range of the Inter-governmental Panel on Climate Change (IPCC) Fifth Assessment Report (AR5) Representative Concentration Pathways (RCPs) as policy scenarios. In order to detect climate variability and change signals, the observed seasonal climate data were subjected to empirical analyses. This involved determination of the first upto fourth moments. The shift in the first moment constituted the trend whose significance was evaluated using the Mann-Kendall test. The moments of both standardized climate data and banana yields were determined and used to identify linkages between current climate variability and banana yields. The Crop Water Assessment Tool (CROPWAT) was used to determine banana water requirements, moisture deficits and yield reductions for the current period. The Pearson’s product-moment correlation coefficient, Refined Wilmott Index and Root Mean Square Errors (RMSE) were used to assess climatemodel performance. Empirical Orthogonal Functions (EOFs) were used to characterize modes of climate variability in both observed and model seasonal rainfall. Comparative graphical analysis based on geo-spatial mapping techniques was used to analyse and map climate variability and change patterns from the high resolution future climate projection information for rainfall, temperature and soil moisture content based on different scenarios. The response of banana growth to expected changes in temperature under vi A1B and A2 scenarios was assessed using a banana-temperature growth regression model. The FAO Eco-crop tool was used to estimate suitable climate conditions for optimal banana growth. The mapping of future suitability of banana during the period (2041-2080) was undertaken using ARCGIS. The results showed that inter annual seasonal rainfall and temperature trends varied between -0.18 to 0.26 mm per year and 0.05 to 0.63 oC per year respectively across the seasons. While there are significant increasing trends in temperature for all seasons at most stations of Uganda, the trends in seasonal rainfall were significant only in a few stations. Further analysis observed significant linkages in variations of current banana yields and climate variability especially with respect to temperature trends. The effect of climate on yields was observed to vary from region to region. This was attributable to variations in other non-climatic factors such as soil fertility and composition, pests and diseases and crop management practices across regions. Results on climate model performance indicated a good match between PRECIS model outputs with observations over most parts of Uganda particularly during October-December. This is mainly attributable to the good representation of the large scale oceanic and atmospheric circulation systems that drive Uganda’s short rains. Results based on Taylor diagrams observed high inter-model variability across different seasons and sub-regions especially during the long (March to May) rain season. Strong coherence in models was evident during the short (October to December) rains. Empirical orthogonal functions analysis also revealed that during October to December, the first mode explained 74.6% and 76.8% of observed and model simulated rainfall variability respectively. The results of climate projections revealed that enhanced rainfall over Uganda is expected under the A1B scenario with depressed rainfall expected under the A2 scenario. In addition, the A1B scenario is projected to exhibit relatively cooler temperatures while the A2 scenario is projected to exhibit relatively warmer temperatures. Consequently, higher soil moisture stresses to banana production are expected under the A2 scenario. Climate extremes that may cause floods and droughts are expected under both scenarios. Comparatively, RCP 2.6 indicated cooler and drier seasonal temperature and rainfall respectively than current observations. Seasonal temperature and rainfall simulation of RCP 4.5 are slightly warmer and wetter than RCP 2.6 simulations. The projected seasonal rainfall is more enhanced in RCP 6.0 vii compared with projections in all other scenarios while RCP 8.5 is associated with highest temperatures for most parts of Uganda. The results of the study further observed that the growth patterns and production of bananas are highly likely to be affected by projected rainfall extremes, increasing temperature and the resultant soil moisture variations across all scenarios and regions. The study observed a larger (smaller) area suitable for banana production under RCP 2.6 and RCP 6.0 (RCP 4.5 and RCP 8.5) for the intermediate period 2041-2060 over Uganda. Due to projected temperature increases across all scenarios, the areas suitable for banana production is likely to reduce under all the four RCP scenarios in the period 2061-2080 relative to the period 2041-2060. It is also expected that the projected temperature changes under the A1B scenario will enhance banana growth. In contrast, banana growth rate under the A2 scenario is expected to decline due to projected warmer temperatures and depressed rainfall under this scenario. The study provides critical science-based evidence of climate variability and climate change over Uganda. It contributes to the understanding of the linkages between banana productivity and observed climate patterns and provides information on the future suitability of the banana crop production under different climate change scenarios. The results from the study are, therefore, pointers for the development of coping and adaptation strategies to expected climate extremes to improve banana productivity, promote the incomes of farmers and enhance food security for sustainable development of Uganda and neighboring countries.