Water storage changes and climate variability within the Nile Basin between 2002 and 2011

TLDR: In this paper, Independent Component Analysis (ICA) was employed to extract statistically independent water storage patterns over the sub-basins from GRACE and the Global Land Data Assimilation System (GLDAS) model.

About: This article is published in Advances in Water Resources.The article was published on 2014-11-01 and is currently open access. It has received 99 citations till now. The article focuses on the topics: Indian Ocean Dipole & Precipitation.

Figures

Table 1: Characteristics of the Nile sub-basins

Table 2: Correlation coefficient between independent patterns of TWS derived from GRACE and climate variability of IOD (in bold) and ENSO (in normal text) over the Nile sub-basins. Correlations are computed at 95% level of confidence. The insignificant values are morked.

Figure 4: ICA decomposition of GRACE TWS signal within the Nile Basin using the proposed ICA method of Forootan and Kusche (2012, 2013). The upper panels are the spatial pattern of each IC. Note that the Nile Basin’s TWS signal is separated (localized) by the ICA method into 4 significant spatially independent components (BEG (left), EH (second from left), LVB (third from left) and Red Sea (right). Together, these four signals computed over the Nile Basin account for 92% of cumulative total variance of TWS changes. In the lower panel, their corresponding time series (ICs) overlaid on smoothed time series of the loading, as well as ENSO and IOD indices.

Figure 5: Dominant independent patterns of GRACE TWS after removing the contribution of Lakes Tana and Victoria surface water signals. Spatial patterns show the concentration of the remaining signals (after removal) over the Ethiopian Highlands (panel A) and the western part of Lake Victoria Basin (panel B). Panels C and D shows their corresponding time series (ICs) overlaid on smoothed time series of the loading and ENSO and IOD

Figure 6: Dominant independent pattern of GRACE-TWS changes for the Nasser region. A) Spatial pattern of IC4 derived from GRACE-TWS changes after correction for the water storage changes of the Red Sea; B) the corresponding temporal evolution of (A) and its comparison with the ENSO and IOD indices; C) comparison of IC4 in (B) with the signal of the Red Sea and Lake Nasser; and D) an overview of IC4 after removing the signal of the Red Sea and its comparison with the signal of Lake Nasser. We should mention here that the Lake Nasser TWS of (C) and (D) represent the same quantity. To enhance the visual comparison, that of (D), is however, vertically shifted to the mean of IC4

Figure 3: (top) Root mean squares of the introduced water storage changes over the major lakes of the Nile Basin, soil moisture changes of the basin and non-tidal water storage changes of the Red Sea; (Bottom) The first three independent modes, derived from ICA. The results show that the strong signal of Lake Victoria (IC3) can successfully be separated from those of the Red Sea and Lake Nasser (IC1) and those of soil moisture (IC2).

Frequently asked questions

Q1. What are the contributions mentioned in the paper "A summary of the paper: “water storage changes and climate variability within the nile basin between 2002-2011”, advances in water resources, doi:10.1016/j.advwatres.2014.06.010" ?

The Gravity Recovery And Climate Experiment ( GRACE ) satellite mission provides a unique opportunity to monitor changes in total water storage ( TWS ) of large river basins such as the Nile. To mitigate this problem, this study employed Independent Component Analysis ( ICA ) to extract statistically independent TWS patterns over the sub-basins from GRACE products. Ethiopian Highlands ( EH ) generally exhibited a declining trend in the annual rainfall over the study period. TWS changes over Bar-el-Ghazal experienced mixed increase-decrease, with ENSO being the dominant climate variability in the region during the study period. 

Q2. What is the effect of the removal of the dominant lake?

The removal of the dominant Lake Victoria signal shows the western side of Lake Victoria having increased TWS, a possible consequence of the long rains of the MAM season that pounds the western side of the lake more. 

Q3. Why is the Nile Basin a difficult place to monitor?

Due to its large spatial extent, however, changes in Nile Basin’s stored water cannot be monitored using traditional methods (e.g., piezometric-based). 

Q4. What is the effect of the removal of the dominant signal of the Red Sea on the TWS?

After the removal of the dominant signal of the Red Sea, the resulting signal (c.f. Fig. 6A) indicates a decline in stored water in the Western Plateau within the Nubian Aquifer covering Lake Nasser at a rate of 2.6 mm/year. 

Q5. What is the smallest change in the TWS over the Ethiopian Highlands?

Analysing GRACE-TWS changes over Ethiopian Highlands (EH) (Fig. 4, IC2) showed a decline at a rate of 18.4 mm/year between 2002-2006. 

Q6. What was the data used in this study?

FIGURE 1The data used in this study consisted of remotely sensed GRACE-TWS changes from 2002 to 2011, Tropical Rainfall Measuring Mission (TRMM)derived precipitation, and water storage data from Global Land Data Assimilation System (GLDAS) hydrological model over the same period. 

Q7. What is the correlation between the TWS changes and the TRMM-rainfall?

This could imply reduced evaporation since the correlation between GRACE-TWS changes and the TRMM-rainfall over BEG gives a phase lag of 1-month at a maximum correlation of 0.53. 

Q8. What is the correlation between IOD and ENSO?

Without the Red sea and Lake Tana’s signals, correlations between IOD and ENSO on the one hand and GRACE-TWS changes on the other hand were respectively 0.33 and 0.30, thus following closely to the LVB pattern (i.e., 0.48 and 0.46 for IOD and ENSO, respectively, for the same). 

Q9. What is the effect of ENSO on the water quality?

ENSO is reported to have a reverse effect, i.e., deficient rainfall tends to occur during ENSO summers (e.g., Eltahir , 1996; Korecha and Barnston , 2007).