A spatial evaluation of multifunctional Ecosystem Service networks using Principal Component Analysis: A case of study in Turin, Italy
The multifunctional Ecosystem Service supply analysis at the spatial level is often the output of a weighted sum of layers in a Geographic Information System (GIS). This procedure is weak in detecting and representing the relationships between the input layers. Nonetheless, composite indicators prod...
Enregistré dans:
| Auteurs principaux: | , |
|---|---|
| Format: | article |
| Langue: | EN |
| Publié: |
Elsevier
2021
|
| Sujets: | |
| Accès en ligne: | https://doaj.org/article/65a42be2f13e4f0fb5bb810c50ece834 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| Résumé: | The multifunctional Ecosystem Service supply analysis at the spatial level is often the output of a weighted sum of layers in a Geographic Information System (GIS). This procedure is weak in detecting and representing the relationships between the input layers. Nonetheless, composite indicators produced by overlaying techniques are quite common in applied research and their discrepancies are underestimated in the scientific community, thus affecting the quality of resulting composite maps. In this work, we empirically test the effectiveness of multivariate statistics to obtain reliable composite Ecosystem Maps in the Turin metropolitan area (north-west Italy). We apply the Principal Component Analysis (PCA, using Matlab and ESRI ArcGis) to seven Ecosystem Service models (Habitat Quality, Carbon Sequestration, Water Yield, Nutrient Retention, Sediment Retention, Crop Production and Crop Pollination) and we evaluate how much the resulting composite map differs from the traditional GIS overlay. In doing this, the spectral analysis (with eigenvectors and eigenvalues) of the covariance matrix of the normalized layers confirms the heuristic arguments about the dependence between Ecosystem Services. We show that the PCA method can provide valuable results in landscape Green Network design, avoiding the limits of standard overlaying procedures. Finally, smoothing and classification techniques, applied to PCA estimates, can further improve the approach and encourage its use in various ecological indicators. |
|---|