Journal of Environmental Hydrology
ISSN 1058-3912

Electronic Journal of the International Association for Environmental Hydrology

JEH Volume 16 (2008), Paper 36    Posted October 16, 2008

Hongqing Wang1
Ehab A. Meselhe1
Michael G. Waldon2
Donatto Surratt3
Sherif Abdou1
Chunfang Chen4
Matthew C. Harwell2

1Center for Louisiana Inland Waters Studies, Institute of Coastal Ecology and Engineering, University of Louisiana at Lafayette, Lafayette, LA, USA
2Everglades Program Team, A.R.M. Loxahatchee National Wildlife Refuge, Boynton Beach, FL, USA
3Everglades Program Team, Everglades National Park, Boynton Beach, FL, USA
4School of Natural Resources and Environment, University of Florida, Gainesville, FL, USA

A water-quality model application to large geographical areas typically involves the spatial aggregation of sub-areas into compartments which are assumed within the model to be uniform. Model compartments often are delineated using available data and professional judgment. Compartment delineation affects the accuracy of hydrodynamic and water quality modeling by subjectively grouping sub-areas of similar characteristics. In this research, we applied cluster analysis (CA) to objectively determine the number of compartments and to spatially delineate compartments with similar descriptive features for water-quality modeling. Here, surface-water quality data collected in the A.R.M. Loxahatchee National Wildlife Refuge (Refuge), Florida were analyzed using CA of concentrations of chloride, total phosphorus, sulfate, and calcium measured at sites distributed throughout the Refuge. The Refuge is a remnant of the northern Everglades wetland ecosystem encircled by a levee and borrow canal. Cluster analysis classified the marsh into six spatial compartments: Perimeter East; Perimeter West; Transition East; Transition West; Interior North; and Interior South. Although the perimeter canals surrounding the Refuge marsh are all connected, they clustered into three water quality compartments: Canal East; Canal Northwest; and Canal Southwest. Cluster validation criteria (pseudo F statistic and the cubic clustering criterion) indicate that there is an advantage of this new compartmental design when compared with our previous compartmentalization method that delineated boundaries using only distance from the canal to the marsh interior and professional judgment. This new approach provides a more technically rigorous compartmentalization for the purposes of water quality modeling. The approach supports more efficient modeling of spatial and temporal variation in water quality along the gradient from both the peripheral canal to the marsh interior, and from west to east and from north to south gradients.

Reference: Wang, H., E.A. Meselhe, M.G. Waldon, D. Surratt, S. Abdou, C. Chen, and M.C. Harwell. 2008. Compartment delineation for a wetland water quality model in the Northern Everglades, Florida, USA. Journal of Environmental Hydrology, Vol. 16, Paper 36.
Hongqing Wang
Center for Louisiana Inland Waters Studies
Institute of Coastal Ecology and Engineering
University of Louisiana at Lafayette
P.O. Box 42291
Lafayette, LA 70504


Return to JEH 2008 Papers

Return to HydroWeb Homepage