6. Discussion#
In this chapter, the implication of the key findings will be discussed. Also, the limitations of the research will be discussed.
6.1 Implication of key findings#
The results for climate scenarios SSP2-4.5, SSP3-7.0 and SSP5-8.5 show that droughts in the Okavango delta become more frequent compared to historical observations. This will have a negative impact on life in and around the Okavango Delta. Cases like hippos that get stuck in mud pools will become more frequent. It also affects the people that are dependent on the Okavango Delta for their livelihoods. The people use the water of the delta for their drinking water, for fishery and eco-tourism is an income source for the region.
For the SSP1-2.6, it is projected that the number of droughts between 2065 and 2099 will be slightly lower than the observed number of droughts from the last 35 observed years. The reason for this could be that in this scenario global CO2 emissions are reduced and will reach net-zero after 2050. In the other scenarios, the CO2 emissions will not become net-zero before 2100. Reducing the CO2 emissions and reaching net-zero by 2050 will lead to a lower concentration of CO2 in the atmosphere. This will result in lower temperature rise for the SSP1-2.6 scenario compared to the other scenario. Because of lower temperatures, the potential evaporation will also lower. This will result that more of the precipitation in the upper catchment area will actually reach the Okavango Delta as less water evaporates.
About 95% of water flowing through the Okavango River is rainwater (Andersson, et al., 2003). The R2 value of 0.46 in figure 5.3 suggests that changes in yearly precipitation alone do not fully explain the changes in yearly volume, although rainfall being the most dominant source of water supply for the Okavango Delta. Although the majority of the flow comes from rainfall, other hydrological processes, like evaporation, also have an important role in how much water actually arrives in the Okavango Delta. So, the number of droughts in the Okavango delta in the future cannot be retrieved from the projected precipitation alone.
6.2 Limitations#
To determine the accuracy of the results, limitations of the research should be considered. One of the limitations is the method of determining threshold values for hydrological droughts in the Okavango River. Connecting historical events with the discharge data has its limitations. First, the used observed data in this study had a timeframe of 46 years and stopped in 2020. So, the case of hippos getting stuck in mud pools in 2024 could not be connected to the discharge data, which could have shown a different result. Also, news articles about droughts in the Okavango Delta are sparse, especially in earlier years like the eighties. Furthermore, although the Okavango River supplies the majority of the water of the Okavango Delta, there are other aspects that influence hydrological droughts, like local precipitation and evaporation. This study uses percentiles to determine the threshold for the different categories of droughts. These threshold values are not based on ecological tipping points. Research on how the area of the floodplains change due to the inflow of the river and how much inflow is needed to fill up the lakes can give threshold values with more ecological meaning.
Another limitation is the assumption that the CMIP6 forcing data for the future has the same bias as the historical CMIP6 forcing data. In this study quantile mapping was used for bias correction, but other methods could also be used. This also can influence the results.
The different climate scenarios and associated CMIP6 forcing data also have limitations. First, not one scenario will be exactly followed. It is most likely that the future path will fall somewhere between the different scenarios. Also, there is uncertainty in the future forcing data as it is impossible to predict the exact weather patterns over a longer period.
Lastly, the HBV model was used to model the Okavango River. The parameters were calibrated by using two thousand different sets and the results were tested against the observed data of 46 years. Although the calibrated parameters in this study showed good results, using more different parameter sets and a longer period of observed data could lead to other parameters and different results. Furthermore, a model does not represent reality because of its simplification. Due to the simplification the results have uncertainty.