A Generalized Model for Estimating the Energy Density of Invertebrates

Invertebrate energy-density (ED) is used in many ecological studies and measured using bomb calorimetry. However, literature ED values are often used, which may not account for spatial and temporal variability, thus decreasing their reliability. We developed and tested a least squares regression model to predict ED from dry weight (DW) proportion based on taxonomically, spatially, and temporally diverse invertebrate samples. The samples included invertebrates from aquatic and terrestrial habitats of four continents and two oceans. A significant relationship between ED and DW proportion (P<0.0001; r²=0.96) was observed in the model (ED (J/g wet weight) = 22,960∙DW–174.2). Model evaluation showed that 98.8% of the variability between observed and predicted values was attributed to residual error and that the 97.5% joint confidence region included the intercept of 0 (–103.0 ⁺/- 707.9) and slope of 1 (1.01 ⁺/- 0.12). This empirical model provides for accurate estimation of invertebrate ED from DW proportion, without the need of bomb calorimetry. Use of this model will increase the reliability of invertebrate ED values compared with using literature values and reduce potential error in models. Because our general model is affected little by taxonomic, seasonal, or spatial variability, it should prove useful for a wide range of ecological studies.