Relationships between tree diversity and aboveground biomass in some hardwood forests of Alabama

Background/Question/Methods

Grassland experiments showed conclusively that the greater the plant diversity, the greater the plant productivity. However, there is lack of such evidence for forests. The crucial research question we want to answer is whether more diverse forests are more productive in terms of aboveground organic matter accumulation. We investigated the nature of the relationship between the predictors tree diversity (Shannon’s index, ShI) and species richness (SpR) and the response variable above ground biomass (AGB). We used publicly available Forest Inventory and Analysis data. The data consisted of all 114 plots from the most recent 2010 annual inventory in Alabama. We chose plots that contained mixed hardwoods and had not experienced disturbance since the previous measurement five years ago. Plot and subplot level relative density was estimated by comparing the stand density index (SDI) to the maximum SDI. Shannon’s diversity index of the stands was estimated based on basal area of the trees. We related the overall plot diversity to the AGB. Additionally, for every plot we compared the AGB of the subplot that had the greatest diversity with the AGB of the subplot that had the smallest diversity.

Results/Conclusions

Estimates of the mean AGB were significantly different for the different levels of relative stand density and for different Shannon’s diversity index classes (p<0.001). The range of total AGB (±standard error) on the study plots was from 1.5±0.7 to 476.2±0.14 ton/ac and the rate of growth was from 0.19 ton/ac/yr to 11.34 ton/ac/yr. The range of Shannon’s index was from 0 to 3.01.There were a total of 97 tree species and species richness ranged from 1 to 38 species per plot. We found significant positive linear relationships between AGB and tree diversity (AGB = -48.108+80.213*ShI; R²=0.50, p<0.001) and between AGB and species richness (AGB = -23.257+7.489*SpR; R²=0.64, p<0.001). However, there was no relationship between the difference in AGB between the subplots with smallest and greatest AGB, and the predictor variable difference in ShI between the subplots with smallest and greatest ShI (p=0.098). We will add additional plots to the data and will choose only stands that have not been cut since establishment and will examine if this may change the observed lack of relationship.