Validity and sensitivity of a model for assessment of impacts of river floodplain reconstruction on protected and endangered species

A statistically significant correlation (p < 0.01) between predicted and observed values for protected and endangered species was found. The sensitivity of the model to value assignment proved to be low. Comparison of five realistic valuation options showed that different rankings of scenarios predominantly occur when valuation criteria are left out of the assessment. Based on these results we conclude that linking species to ecotopes can be used for adequate impact assessments. Quantification of sensitivity of impact assessment to value assignment shows that a model like BIO-SAFE is relatively insensitive to assignment of values to different policy and legislation based criteria. Arbitrariness of the value assignment therefore has a very limited effect on assessment outcomes. However, the decision to include valuation criteria or not is very important.