Performance of three rainfall interception models with variable canopy cover fraction (c) and canopy storage capacity (S), from satellite-based leaf area index (LAI) data.

The most common rainfall interception models (Rutter, Gash and Liu) require the knowledge of two canopy-related parameters, the canopy storage capacity (S) and the canopy cover fraction (c). Even though canopy cover changes over time, these parameters are treated as constants in most rainfall interception studies. The aim of this study is to evaluate the performance of these three interception models with the use of time-variable S and c with meteorological and throughfall data from a semi-arid Pinus brutia forest (Cyprus). Leaf area index (LAI) were acquired from the Copernicus global land service (https://land.copernicus.eu/global/products/lai). These data were interpolated with a cubic spline function to obtain a daily time series. Daily S and c values were expressed as one-parameter linear (S) and exponential (c) functions of the daily LAI values The model results with the variable S and c were compared with the model calibration and validation results obtained with constant S and c values. The interception losses computed with the three models ranged between 18 and 20% of the total rainfall. 
All three models showed high performance for both calibration and validation periods with Kling–Gupta Efficiency (KGE) above 0.90. However, the constant S and c models show equifinality, meaning that a range of combinations of the input parameters S and c will result in the same interception loss. The Gash model with the variable S and c resulted in higher KGE (0.968) and lower percent bias (0.8%) than the Gash model with constant S and c (0.956 KGE and 1.5% percent bias), during the calibration period. Rutter and Liu models with the variable S and c resulted in lower bias (-6 mm and -11 mm) than the models with constant S and c (17 mm and 27 mm). The models were all capable of capturing the inherently variable interception process. However, ground-based LAI data are needed to validate the satellite-based data. 
This research has received funding from the European Union's Horizon 2020 Research and Innovation programme, under Grant Agreement 641739 (BINGO Project) and from the Research and Innovation Foundation of Cyprus, through the Water Joint Programming Initiative (FLUXMED project).