START flash tag cloud requires Flash Player 9 or better.Tag Cloud created by Roy Tanck
January 13th, 2011
Previous water budget studies over the Lake Victoria basin have shown that there is near balance between rainfall and evaporation and that the variability of Lake Victoria levels is determined virtually entirely by changes in rainfall since evaporation is nearly constant. It is also well known that the variability of rainfall over East Africa is dominated by ENSO. However, the dipole mode is the second most dominant rainfall climate mode also accounts for signiﬁcant variability across the region. The hydrologic adjustment time of the lake is also on a decadal scale.
Based on this knowledge, we hypothesize that ENSO dominates the variability of Lake Victoria levels. We further hypothesize that the dipole mode and hydrologic adjustment also play signiﬁcant roles but we are uncertain which of these two is more signiﬁcant. The relationship between the ENSO and dipole mode with Lake Victoria levels is nonlinear and a key objective in this study is to estimate the relative contributions of these modes in modulating Lake Victoria levels to test our hypothesis.
We ﬁnd that the sudden signiﬁcant increase in lake levels from 1961-64 was mainly caused by consistent above average precipitation during those years, while the decline from 1965-2005 was due to the lake reaching a new equilibrium level. The ﬁrst EOF mode of annual precipitation variability (ENSO) accounts for the highest impact on the annual variability of Lake Victoria levels. While the second annual EOF mode accounts for approximately 10 percent of the variability, its affect on the variability of lake levels is nearly negligible because the loadings are very small over the lake.
Lake Victoria levels could reach 14 meters under IPCC AR4 A2 scenario projections. We map the potential ﬂooding from these increased levels using GIS. This information is potentially highly valuable in assessing future use of hydroelectric dams and other applications such as land use and infrastructure planning over the Lake Victoria basin.
The annual variability in precipitation in terms of the seasonal variability from station gauge precipitation over Uganda, Kenya and Tanzania is investigated by performing seasonal EOF analysis on station gauge precipitation for the October-December (OND), March-May (MAM), January-February (JF) and June-September (JJAS) seasons. The EOF time series of the dominant modes are correlated with SST anomalies for each season to help determine their sources of variability. The ﬁrst annual mode is correlated with the ﬁrst modes of OND and MAM seasons, while the second mode is correlated with the second modes of MAM and JJAS seasons.
We then perform EOF analysis on a ten year running mean of 1961-90 OND season station precipitation to investigate decadal variability. Correlation of the time series of the ﬁrst EOF mode with SST anomalies gives a ‘ENSO-like’ signal pattern consistent with recent studies.
We recommend that there is a deﬁnite need for sustained and improved monitoring of the lake and its basin, to compile high quality monthly observations for all the factors in the water balance model including information on water taken from the lake and its tributaries for human use. We further recommend the use of a dynamical hydrological model for use in projections of lake levels for climate change scenarios. Such a model would have more reliability outside of the regimes under which the model used in this study was calibrated. Future studies would beneﬁt from an estimation of the cascade of uncertainty over the entire chain of steps involved from data collected to impact modeling.