How to identify the individual climate (Precipitation, Tmax & Tmin daily data of 15 years period) and basin parameters (Land Use, Soil and Slope) contributing and its relative importance to streamflow using the Boosted Regression Trees in R (or any other model)?
With respect to climate parameters, the comparison can be straightforward made with the streamflow statistically. However, the basin parameters considered (Land Use with 8 classes, Soil with 6 different types and Slope range from 0-86.2°) doesn’t match up with the long time series data of streamflow.
So, considering the above condition, on what basis can the basin parameters along with the climate variables be made to identify the contribution and its relative significance to streamflow.
Any help would be highly appreciated.
Thank you
I would select the period of streamflow data that corresponds to the time period of your basin variables. And then fit them together in a BRT model. You can use the gbm package and caret package in R to fit your models. In your output, your explanatory variables (climate and basin variables) will be ordered according to the importance in explaining your streamflow data.
See here for a good example: https://www.storybench.org/tidytuesday-bike-rentals-part-2-modeling-with-gradient-boosting-machine/