Non-destructive biomass estimation for eelgrass (Zostera marina): Allometric and percent cover-biomass relationships vary with environmental conditions

Abstract

Estimating plant biomass reliably over large areas while minimizing impacts on sampled habitats is an important goal in plant ecology. Often, this is accomplished by first using a small number of harvested plants to quantify the relationship between plant biomass and less destructive predictor variables (e.g., height, cover), and then applying this relationship across larger spatial scales. However, the influence of environmental conditions on these relationships is often poorly understood. Here, we assess the impact of environmental variability on two biomass estimation functions for the seagrass Zostera marina in Atlantic Canada: the allometric leaf length-weight relationship and the relationship between percent cover and above-ground biomass (AGBM). First, we determined allometric and cover-AGBM regression relationships at the regional level using data from all sites pooled. We then investigated whether these models could be improved by including a site group covariate based on principal component analysis of site-level environmental data. At the regional level, allometric and cover-biomass models were both strongly significant, although uncertainty was high in the cover-AGBM model. Both models improved markedly when environmental variability (i.e., site group) was included: in warm, shallow conditions, eelgrass leaves were lighter for a given length, and AGBM increased at a slower curvilinear rate with percent cover. This indicates that environmental effects on eelgrass morphological traits not typically included in biomass models (e.g., leaf thickness, rigidity) can be important. Our study suggests that environmental effects on eelgrass biomass models should be considered, particularly when highly accurate estimates with low uncertainty are required.