Supplementary material 1 (S1) and 2 (S2): Predicting soil moisture conditions across a heterogeneous boreal catchment using terrain indices

S1: DEM preparation and terrain indice modeling, Python Code and DEM

This folder contains all data needed to produce the terrain indices modeled within the paper.
It contains the Python Code for DEM aggregation, hydrological correction and terrain indice modeling

S2: Interactive OPLS analysis loading plot for the Krycklan catchment
and DEM-derived terrain indices in respect of soil moisture predictions.

This folder contains the Interactive version of the OPLS analysis loading plot (Fig 2.)
for the Krycklan catchment and DEM-derived terrain indices in respect of soil moisture predictions.
In the interactive version of this graph (html), all labels are visible by moving the cursor on top of each circle.
By clicking on a circle the terrain chosen index group will be highlighted

Variables that cluster closely within the same neighborhood along the far sides
of the horizontal axis are the more robust soil moisture predictors across DEM
scales. Colored guides connect terrain indices moving from small to large resolutions
as depicted by the symbol size. In the loading plot, predictive performance increases
with increased distance from 0 on the predictive axis (pq[1]). Negative and positive values
on the (pq[1]) axis correspond to negative and positive correlations with Y. The orthogonal axis
(poso[1]) represents how much of the variation for each variable was not correlated with the determinant (Y).