Niva biodiversity index
Niva's biodiversity index is a potential of biodiversity based on landscape metrics. To make this index more concrete for users a simplification using decision tree's has been derivated from the statistical model. We use CONTRA model (LINK TO PAPER) to build the decision tree. Depending on the number of landscape metrics available the index can have different levels of precision from Tier 1 (lowest precision) to Tier 3 (highest precision). Required data (as geographical layers) and computed metrics for each Tier ared dispalyed below. We also plan to to provide methods to predict Niva's biodiversity index at different scale: on grid of \(1km^2\), at farm level (not done yet).
As the model is calibrated on a dataset and cannot be applied on landscapes that have features too far from the calibration dataset a novelty index based on mahalanobis distance is computed alongside the biodiversity index to provide prediction's confiance. For each Tier there is a threshold to consider prediction as valid.
For crops diversity Niva provide a specific nomenclature to group crops.
Tier 1¶
Required layers¶
| Layer | Description |
|---|---|
| IACS | Integrated Administration and Control System with crop plots. |
| ARTIFICIAL | Elements that are impermeable (buildings, roads, train lines etc). |
| NAA | Non Agricultural Areas such as trees, vegetation, etc.. |
Landscapes features¶
| Landscape feature | Short Name | Description |
|---|---|---|
| Semi Natural Cover | SNC | Cover (%) of semi-natural elements |
| Mean Field Size | MFS | Mean of crop fields area. |
| Richness | RICH | Number of different crops. |
| Diversity | DIV | Inverse of simpson index on crops types. Represent diversity of crop mosaic. |
Novelty threshold¶
A sample with Novelty index above 1.25 is considered as too far from the the calibration set and not valid.
Processing algorithms¶
Tier 2¶
Not done yet.
Tier 3¶
Not done yet.