Predicting EC for Drainage Water Management

Abstract

In 1982, 405 measurements of the subsurface drainage water quality parameter electrical conductivity (EC) in dS/m were collected. The data were obtained from 15 subsurface drains on 9.3 ha (23 acres) of irrigated agricultural land at Fallon, Nevada. The temporal and spatial variabilities of EC were studied using time series concepts. Autocorrelation functions (ACF) were used to evaluate the magnitude of temporal and spatial variations of EC. Results suggest that an 11‐week period is the maximum time interval for sampling and that the sampling spatial interval of 36.6 m (120 ft) is too large. Autoregressive integrated moving average (ARIMA) models for both temporal and spatial structures were established through the Box‐Jenkins time‐domain modeling process. The degree of uncertainty of the forecasts were tested using after‐thefact forecast procedures. These models can be used for various purposes such as forecasting future values and determining the transfer function, which provides a way to relate water management plans to water quality control.