Predictive modeling of floristic components using soil variables in seasonal lacustrine wetlands

Abstract

The overlap between vegetation and environmental factors in wetlands has always been a key area in plant ecology. Studying vegetation components in wetland without allusion to the environmental factors governing their distribution is inconclusive. To this end, mathematical models were applied to predict floristic components using soil variables in seasonal lacustrine wetlands. A quadrat size of 5m×5m was used to systematically sample the vegetation for density, frequency, height, basal area and crown cover. In each quadrat, soil samples were obtained at different rooting depths (0 -15cm and 15–30cm). Results obtained showed variations in the floristic composition in the wetlands seasonally. Rural wetland had seventeen (17) plant species during the dry season and fourteen (14) plant species during the wet season while urban wetland had eleven (11) plant species in the dry season and thirteen plant species (13) in the wet season. Variations in density, height, frequency, basal area and crown cover occurred in the wetlands. The physicochemical properties of the soil in these wetlands varied seasonally. A prediction model using stepwise multiple regression analysis revealed the vegetation and species responses to the environment on the basis of soil variables and also showed their relationship with each other. It also showed the various predictors (soil properties) of the vegetation components such as density (exchangeable acidity, sand, pH, Na, Zn, Cd, Ca and total nitrogen), height (base saturation, Ni and sand), crown cover (exchangeable acidity, clay and electrical conductivity) and basal area (exchangeable acidity). In summary, it is established that using mathematical models, environmental variables (soils) can serve as good predictors of vegetation components in wetlands by elucidating the soil-vegetation interrelationships.