A validation of the model for effluent disposal using land irrigation (MEDLI)

Biological waste produced by intensive livestock farming is a valuable and useful product used in the agricultural industry for irrigation of crops. Manure and liquid effluent contains nutrients that can be effectively utilised in commercial cropping of livestock feed. It provides farmers with a low cost organic material and fertilizer, resulting in high yielding crops if carefully managed. The downside of using effluent in crop irrigation is the potentially high concentrations of chemicals and pathogens in effluent can, if mismanaged, reach toxic levels in the soil. This can lead to crop failure and in worst cases, land and water contamination. It is a requirement of Queensland law that before an enterprise irrigates with effluent it must first obtain a regulation certificate. To fulfil this obligation an effluent irrigation scheme must be designed and modelled to the satisfaction of the regulatory authority. There are tools available which aid the designer of the effluent irrigation scheme in conducting water and nutrient balances. The software package recommended by the Queensland Government is; The Model for Effluent Disposal using Land Irrigation (MEDLI). The purpose of this program is to model; effluent volumes, concentrations of chemical constituents in effluent, point of deposition soil chemistry and nutrient uptake by plants. Due to the absence of previously completed program validation, this research aimed to conduct validation of MEDLI software. Modelling scenarios were entered into the program using input variables that had been established from data collected from three beef cattle feedlots. Scenarios were set-up to try and best mimic site conditions, so a comparisons could be drawn between the simulated and observed datasets. Results of the comparisons for all three sites found, often significant variation in the values of simulated and observed conditions. Weak correlation of the datasets could not be conclusively attributed to systematic errors in the model. Analytical errors such as; improperly defined inputs and inadequacy of sample sizing may have contributed to the bias found between datasets. A particularly notable conclusion of the analysis was that far greater definition is required around the required estimations of the pre-treatment and anaerobic pond chemistry inputs. A recommendation is; MEDLI literature which is supplied with the program, should provide considerably more detailed guidance on deriving accurate estimation of these input variables