You are here:
Assessing Wildlife Habitat Value of New England Salt Marshes: II. Model Testing and Validation
Citation:
MCKINNEY, R. A., M. A. CHARPENTIER, AND C. WIGAND. Assessing Wildlife Habitat Value of New England Salt Marshes: II. Model Testing and Validation. ENVIRONMENTAL MONITORING AND ASSESSMENT. Springer, New York, NY, 154(1-4):361-371, (2009).
Impact/Purpose:
This paper tests an assessment model to quantify the wildlife habitat value of New England salt marshes with bird abundance and species richness data from sixteen salt marshes in Narragansett Bay, Rhode Island. The paper also describes the use of an Akaike information criteria modeling approach to optimize the assessment for a group of species (wading birds of the family Ardeidae) that are important users of coastal salt marshes in New England. The assessment model can be used to generate data on wildlife habitat value to supplement comprehensive assessments when assessing areas for which habitat value is of particular importance or concern. In addition to validating the model, our results suggest that relatively small wetlands in highly urbanized coastal settings can provide important wildlife habitat value if key habitat attributes are present.
Description:
We test a previously described model to assess the wildlife habitat value of New England salt marshes by comparing modeled habitat values and scores with bird abundance and species richness at sixteen salt marshes in Narragansett Bay, Rhode Island USA. Assessment scores ranged from 37 – 61 % of the maximum attainable score, and 11 of the 16 marshes scored above the 50th percentile, indicating that the majority of our study marshes provide significant habitat value to terrestrial wildlife. We recorded 31 bird species at our sites, including 6 members of the family Ardeidae (herons, egrets, and bitterns). Species richness (r2 = 0.24, F = 4.53, p = 0.05) and abundance (r2 = 0.26, F = 5.00, p = 0.04) of Ardeids significantly increased with increasing assessment score. We optimized our assessment model for Ardeids by using Akaike information criteria (AIC) to compare a series of models comprised of specific components and categories of our model that best reflect their habitat use. The model incorporating pre-classification, wading bird habitat categories, and natural land surrounding the sites was substantially supported by AIC analysis as the best model. The abundance of Ardeids significantly increased with increasing assessment scores generated with the optimized model (r2 = 0.48, F = 12.5, p = 0.003), demonstrating that optimizing models can be helpful in improving the accuracy of the assessment for a given species or species assemblage. In addition to validating the assessment model, our results show that in spite of their urban setting a majority of our study marshes provide substantial wildlife habitat value. This suggests that even small wetlands in highly urbanized coastal settings can provide important wildlife habitat value if key habitat attributes (e.g., natural buffers, habitat heterogeneity) are present.