Detection, measurement and prediction of shoreline recession in Accra, Ghana

Coastal mapping, using various data capture and feature extraction techniques, has furthered understanding of trends in shoreline evolution by allowing calculation of accurate historic rates of change that subsequently enable the prediction of future shoreline positions through different modelling procedures. The results have helped influence coastal policy formulation and promoted the development of sustainable management practices in coastal regions throughout the developed world. However, sustainable coastal management is rarely practiced in developing countries, one of the fundamental reasons for this being a general lack of reliable and accurate historic data on shoreline position. Previous studies on the Ghanaian coastal region of Accra, where accurate and reliable geospatial data for analysis is scarce, have reported erosion rates of anything between two and eight metres per year. This high level of inconsistency in reported rates has hindered effective and sustainable coastal management. The research reported in this paper addresses this issue, using mapping data from 1904, 1974, 1996 and 2002 to estimate, by linear regression, shoreline recession in the Accra region. Predictions for the next 250 yr were then undertaken using a variety of techniques ranging from a process-based numerical model, SCAPE, to geometric approaches including historical trend analysis, the modified Bruun model and Sunamura’s shore platform model. Uncertainties in the various input data were accounted for, including historic recession rates, rock strength, sediment content and, importantly, future sea-level rise under different climate change scenarios. The mean historic rate of erosion in the Accra region was found to be 1.13 m/yr(±0.17 m/yr), significantly less than previously reported, though still very high. Subsequent predictions were used to identify a series of significant economic, ecological and social features at risk, and to estimate when they will most likely be lost to erosion if left unprotected. The case study illustrates that, provided suitable predictive models are selected and the uncertainties involved in working with limited data sets are dealt with appropriately, it is possible to provide statistical information in support of sustainable coastal management for developing countries in the face of a changing climate.