Evaluation of modified DRAINMOD in predicting groundwater table fluctuations and yield of canola in paddy fields under snowy conditions (case study: Rasht Iran)

The computer simulation hydrologic model DRAINMOD and modified DRAINMOD (Luo et al. 2000) for cold condition was used to predict the performance of surface drainage and yield of canola as post‐rice cultivation during the wet growing season of 2004–2005 in Rasht Iran. Performance of surface drainage system treatments with 2 m spacing and 15 cm drain depth as compared with plots having no drainage (4 × 10 m2) were used to evaluate the models. Evaluation of predicted daily groundwater table depths as compared with measured values shows that the root mean square error (RMSE) was about 10.2 and 9.9 cm for DRAINMOD versus 11.0 and 11.5 for modified DRAINMOD in 2 m drain spacing and no drainage treatment respectively. Original DRAINMOD and modified DRAINMOD predictions for groundwater table depth was on average about 4 and 7% less than measured values for the surface drainage treatment and 22 and 14% greater for the no drainage treatment whereas the standard error of original DRAINMOD was 1.4 and 1.51 (cm) respectively. The average absolute deviations were 8.67 and 9.22 (cm). However by using the modified DRAINMOD these values were decreased to 1.34 1.3 8.05 and 8.19 cm respectively. Comparisons show that modifications introduced in DRAINMOD for snowy conditions improved the capability of the model to predict groundwater table fluctuations.