Predictive modelling of eelgrass (Zostera marina) depth limits

Empirical models relating secchi depths to maximum depth limits of eelgrass (Zostera marina L.) can describe basic differences in depth limits between areas or time periods exhibiting large differences in secchi depth. However, these models cannot predict the precise depth limit at a particular site at any specific time. In this study we aim to improve the ability of regression models to predict maximum depth limits by: (1) assuming that eelgrass depth limits respond to changes in secchi depth with a temporal delay of 1–2 years, (2) including other water-quality variables in addition to secchi depth, and (3) taking into account that factors regulating depth limits may vary between years and between sites. We were not able to improve the models by introducing a systematic delay in the response of depth limits to changes in secchi depths. The reason for this failure is likely to have been the systematic nature of our approach, since some sites responded with a delay, while others did not. The explanatory power of the models increased when additional water-quality variables were added in a multiple regression model. Where secchi depth alone explained 58% of the variations in depth limits, addition of winter [NH4+] and maximum water depth as independent variables increased the explanatory power to 71%. These models applied to data from one specific year, but when data from several years (1989–1998) were included, only 35% of the variation in depth limits could be explained by the three factors. More detailed analyses showed that the regulation of eelgrass depth limits varied considerably between years and between sites, and the models were further improved by taking this information into account. Our results confirmed previous studies by showing light to be the most important parameter in the regulation of eelgrass depth limits, but also revealed a complexity in the regulation of depth limits not expressed in earlier studies. Limited colonisation potentials may delay the response to improved light conditions, and hypoxia/anoxia and indirect effects of nutrients may prevent eelgrass from attaining the depth limit that light levels would allow. The power to predict depth limits on the basis of secchi depths can therefore be improved by taking site-specific information on eelgrass growth conditions into account.