Restoring jurisdictional wetland hydrology does not ensure that the hydrologic conditions of any specific natural forested wetland community are recreated. This is especially problematic for a state like North Carolina, which has roughly two dozen different forested wetland communities. Because forested wetland communities align themselves across edaphic and hydrologic gradients, we suggest exploiting these relationships to guide restoration design and set performance standards. However, methods to compare hydrology between multiple reference sites and restored wetlands are lacking. To address this, we studied three different forested wetlands across 13 reference sites to determine which measures of water table levels, based on correlation with plant community composition, would be useful to distinguish these three communities. We then used the best measures of water table levels to assess two restoration sites and compare them with the reference sites. Our results showed that monthly median water table levels encompassing the start of the growing season had the strongest correlation with plant community composition, whereas roughly five other measures also had strong correlations. On the basis of the best measures, both of the restoration sites had water table levels that generally fell within the natural range of the reference sites. Because there was variation in water table levels across each restoration site, the different measures where useful to identify which areas were least similar to the reference sites and might need further monitoring in the future. On the basis of our results, we recommend using community-specific measures of hydrology to guide and assess forested wetland restoration. Copyright © 2013 John Wiley & Sons, Ltd.