In order to account for this, and based on practical
observations, it was hypothesized that there is a level of
persistence in the snow effect data, where the effect on a
given day would be somehow related to the losses of a
previous day, due to the build-up of snowfall on the module.