Refinement of hydrogeochemical models of the ecological impact of acid deposition

Fi-John Chang, Purdue University

Abstract

Four process-oriented watershed acidification models that have been proposed to assess or predict the effects of acid deposition on a watershed were reviewed. These models are the Birkenes, Trickle-Down, MAGIC, and ILWAS models. An examination of each model's concepts, critical assumptions, data requirements, and application results has been performed. The advantages and limitations of each model were discussed. A sensitivity analysis was performed on the ILWAS model to identify and to eliminate the least sensitive parameters and thus to reduce the number of parameters needed for simulation. A common difficulty in applying the watershed acidification models to new watersheds is the problem of parameter estimation. A systematic parameter estimation strategy was developed. It includes a regionalized sensitivity analysis and an automatic parameter calibration technique. The strategy was developed and implemented using the Birkenes and ILWAS models. The regionalized sensitivity analysis is based on a large number of Monte-Carlo simulations to identify the sensitive parameters and to establish a range of values appropriate for each sensitive parameter. The automatic parameter calibration technique adjusts the values of the sensitive parameters based on changes in the residual variances between the predicted and observed streamflow quantity and chemistry. A Golden Section optimization method was used for the minimization of the residual variances. The results show that, using the systematic parameter estimation strategy, significant improvements in the model prediction of the streamflow and its chemistry can be achieved. Two hypotheses--(1) increasing sulfate deposition will increase outflow cation concentration and (2) increasing rainfall intensity will increase outflow hydrogen concentration--were tested with the ILWAS and Birkenes models. The ILWAS model confirms both hypotheses, while the Birkenes model does not produce the same conclusions. This is probably due to the simplicity of the hydrologic submodel of the Birkenes model and its assumption of constancy of the cation exchange capacity. This process-oriented model provides the essential insight to make predictive assessments of short-term and long-term effects of surface and subsurface water acidification under different acid loadings. Application of the ILWAS model to the Walker Branch watershed failed to reproduce the time distribution of the outflow ion concentration fully. (Abstract shortened with permission of author.)

Degree

Ph.D.

Advisors

Delleur, Purdue University.

Subject Area

Civil engineering

COinS