Skip to main content
ARS Home » Research » Publications at this Location » Publication #86035
Title: VALIDATION OF THE HYDROLIC EVEALUATION OF LANDFILL PERFORMANCE (HELP) VERSION 3 MODEL FOR LANDFILL TEST CELLS IN MICHIGAN, UTAH, AND DELAWARE

Author
item Ascough Ii, James
item Rector, Harriet
item DERR-ASCOUGH, LOIS - CSU
item MALTBY, VAN - NCASI

Submitted to: Meeting Abstract
Publication Type: Proceedings
Publication Acceptance Date: 5/21/1997
Publication Date: N/A
Citation: N/A

Interpretive Summary: HELP Version 3 model validation was performed using measured data from landfill test cells in Michigan, Utah, and Delaware. Event-based (monthly & length of simulation (mean annual and cumulative) model output responses were analyzed. Nash-Sutcliffe (NS) model efficiencies for monthly model predictions were generally poor. In general, HELP 2 performed better than HELP 3 for the monthly results as indicated by higher NS model efficiencie The poor monthly results demonstrate that evaluation of potential landfill designs using HELP 3 should be done for long-term water balances only. Analysis of the mean annual and cumulative results showed that the HELP 3 model performed much better as compared to the monthly results. Across all landfill test cells, cumulative relative errors in output response variables ranged from -0.22% to -64.6%. Across all landfill test cells, mean annual relative errors ranged from -0.46 percent to +62.3 percent. HELP 3 model performed much better than the HELP 2 model for the mean annual & cumulative results, as indicated by lower relative errors across all cells & output response variables. Finally, the cumulative results ndicated that the HELP 3 model performed very well considering that no Calibration occurred. The cumulative relative errors across all landfill test cells and output response variables were generally ten percent or less.

Technical Abstract: HELP Version 3 model validation was performed using measured data from landfill test cells in Michigan, Utah, and Delaware. Event-based (monthly and length of simulation (mean annual and cumulative) model output responses were analyzed. The Nash-Sutcliffe (NS) model efficiency statistic was used to analyze the monthly output responses. Relative error and absolute relative error were used to analyze the mean annual and cumulative output responses. HELP Version 2 and HELP 3 monthly, mean annual, and cumulative model output responses were compared using the above statistical methods. The NS model efficiencies calculated from the monthly model output responses were very poor. In general, HELP 2 performed better than HELP 3 for the monthly analysis. Analysis of the mean annual and cumulative output responses showed that HELP 3 performed much better in comparison to the monthly analysis. Furthermore, HELP 3 performed much better than HELP 2 for mean annual and cumulative output responses. The mean annual and cumulative results indicate that the HELP 3 model performed very well considering that no calibration occurred. In general, the cumulative relative errors across all landfill test cells and model output response variables were less than ten percent, with only a few exceptions.