However, the solution is sensitive to the initial guess value and becomes the same as the initial value for grids without any radio ray path passing by.
Station-based observations of monthly temperature and precipitation were mapped on a 0.5° latitude/longitude grid and then used to calculate monthly PDSI values for each grid point (R.R. Heim, Jr., National Climatic Data Center, Ashville, NC, USA) by using climatically aided interpolation techniques (16 ).
Because the simulated results are mean values for each grid, we compared the simulated results with station observations derived from the corresponding grid cells.
Numerical values for the grid work and wear rates are calculated from the predicted forces during the rod impacts against the supports.
The element-wise multiplication of this matrix by the posterior matrix (containing the posterior values for each grid node) ensures the correct numerical evaluation of Eq. (7).
As expected, the relative error in the mean Lichen Biodiversity (Italian acronym: BL) values and the error associated with the interpolation of BL values for (unmeasured) grid cells increased as the sampling density decreased.
In order to assess the significance of the level of variation that was explained by our models, we ran additional models in which the environmental layers were substituted by layers with random values for each grid cell.
The variable combination for the best fit model was then used to predict C values for unsampled grid cells by splitting the sample data into a training set (two-thirds) to build the model and a test set (one third) to evaluate how well the model predicts C for the withheld points.
Furthermore, because no formalized significance testing has yet been developed for GDM, to assess the significance of the level of variation that was explained by our models, we ran additional models in which the environmental layers were substituted by layers with random values for each grid cell.
For all those time series, we estimated mean values A0 as well as annual (A1, B1) and semi-annual (A2, B2) variations within a least-squares adjustment, performed separately for each parameter r(t) (1) The estimated values for each grid point were saved to an external ASCII-file (14 MB at 1°, 0.5 MB at 5° resolution), which must be loaded with every run of the GPT2 subroutine.
