Comparison of specific and universal linear regression models for predicting soil water retention in eastern Croatia

We analyzed 96 samples from 48 soil profiles (Gleysols, Chernozems, Anthrosols) on arable land in eastern Croatia (Europe) in order to develop specific pedotransfer functions (PTFs) for the local-scale prediction of soil water retention at field capacity (FC) and wilting point (WP). Our easy-to-use multiple linear regression models performed very well (RMSEs 2.7% and 1.8%, MAEs 2.1% and 1.5%; for FC and WP, respectively), in spite relying only on the contents of clay, silt, and organic C as predictors. These three are routinely measured, fundamental soil properties for which the data is widely-available. Because these PTFs were developed for the soils on the most common groups of parent materials in continental Croatia (loess derivates and fluvial deposits with mostly silt loam, silt clay loam, and silt clay textures), they should have wide applicability. We have also tested the prediction performance of the universal regression-based PTFs (EU-PTFs), created by Toth et al. (2015) to predict FC and WP at the European scale. Although the soils of Croatia are not included in the EU-HYDI database, which was used for the development of the EU-PTFs, we found that the continental model predicts the FC of our soils almost equally well (RMSE = 3.3%, MAE = 2.7%) as our local model. However, for predicting the WP of our soils, the continental model is not as accurate, overestimating this parameter notably more (RMSE = 4.9%, MAE = 4.4%) than the local model. Given that the WP values of our (and many other European) soils are relatively low, the obtained prediction errors may be even more problematic. However, our results suggest that the use of EU-PTFs can partially compensate for the development of specific (local) PTFs, at least in the European regions with soilscapes similar to those in our study. To help increase the reliability of the EU-PTFs, the soil database used for their development should be expanded to include new countries and regions of Europe.