Modeling the subsidence of peat soils in the Dutch coastal area

In large parts of the western coastal region of the Netherlands peat soils occur. Historic land use resulted in decomposition of organic matter and compaction of these soils and as a consequence the land surface subsided. Along with sea level rise and tectonic subsidence since the end of the Weichselian ice-age, this resulted in a country where one-third of the land lies below mean sea level. Global warming will accelerate the sea level rise and land subsidence due to peat oxidation, leading to elevated risks of flooding and increasing upward seepage of brackish and nutrient rich ground water. In the Polder Groot-Mijdrecht near Amsterdam the land surface is currently between 5 and 6.5 m below sea level. For this polder we calibrated a model for the land surface subsidence. The subsidence rate is modeled using a differential equation accounting for the peat oxidation and other processes causing land surface subsidence. The model input consists of information on (i) the thickness of peat layers derived from soil surveys in 1964 and 1967; (ii) the regulated surface water levels in a network of drainage ditches, and (iii) the initial elevation measured between 1954 and 1968 at 1423 locations. The model was calibrated on the elevation measurements in 2005 and on the peat thickness from a recent soil survey in 2005. The subsidence between 1954 and 2005 as predicted with the model was validated with a subsample of 100 points out of the 1423 locations. The average rate of predicted subsidence between 1954 and 2005 was approximately 5 mm year(-1). In soils containing peat layers the predicted subsidence rate due to peat oxidation was up to 8 mm year(-1). On the contrary, at locations without peat the predicted subsidence rate was only 0.7 mm year(-1). The model accounted for 78% of the variance of the historic subsidence, and the root mean squared error of the predicted elevation in 2005 was 94 mm. Model forecasts of the cumulative subsidence between 2005 and 2020 for the eastern half of the study area where still thick peat layers occur, were over 80 cm. In areas where peat layer in the soil profile are absent the forecasted cumulative subsidence was 10 mm. The case study shows that the simple process model, when calibrated on local data, can help in evaluating flooding risks and in taking long-term strategic decisions on water management and land use in peat areas. Crown Copyright (C) 2011 Published by Elsevier B.V. All rights reserved.