Understanding the decline of water storage across the Ramser-Lake Naivasha using satellite-based methods

Abstract

It has been postulated that Lake Naivasha, Kenya, has experienced a rapid decrease (and fluctuations) in its spatial extent and level between the years 2002 to 2010. Many factors have been advanced to explain this, with horticultural and floricultural activities, as well as climatic change, featuring prominently. This study offers a multi-disciplinary approach based on several different types of space-borne observations to look at the problem bedeviling Lake Naivasha, which is a Ramsar listed wetland of international importance. The data includes: (1) Gravity Recovery and Climate Experiment (GRACE) time-variable gravity field products to derive total water storage (TWS) variations within a region covering the Lakes Naivasha and Victoria basins; (2) precipitation records based on Tropical Rainfall Measurement Mission (TRMM) products to evaluate the impact of climate change; (3) satellite remote sensing (Landsat) images to map shoreline changes and to correlate these changes over time with possible causes; and (4) satellite altimetry observations to assess fluctuations in the lake’s level. In addition, data from an in situ tide gauge and rainfall stations as well as the output from the African Drought Monitor (ADM) model are used to evaluate the results.This study confirms that Lake Naivasha has been steadily declining with the situation being exacerbated from around the year 2000, with water levels falling at a rate of 10.2 cm/year and a shrinkage in area of 1.04 km2/year. GRACE indicates that the catchment area of 4°×4° that includes Lake Naivasha loses water at a rate of 1.6 cm/year for the period from August 2002 to May 2006, and 1.4 cm/year for the longer period of May 2002 to 2010. Examining the ADM outputs also supports our results of GRACE. Between the time periods 2000–2006 and 2006–2010, the lake surface area decreased by 14.43% and 10.85%, respectively, with a corresponding drop in the water level of 192 cm and 138 cm, respectively, over the same periods. Our results show a correlation coefficient value of 0.68 between the quantity of flower production and the lake’s level for the period 2002–2010 at 95% confidence level, indicating the probable impact of anthropogenic activities on the lake’s level drop.