Assessing and improving the transferability of current global spatial prediction models

Ludwig, M; Moreno-Martinez, A; Hölzel, N; Pebesma, E; Meyer, H

Forschungsartikel (Zeitschrift)

Zusammenfassung

Aim: Global-scale maps of the environment are an important source of information for researchers and decision makers. Often, these maps are created by training machine learning algorithms on field-sampled reference data using remote sensing information as predictors. Since field samples are often sparse and clustered in geographic space, model prediction requires a transfer of the trained model to regions where no reference data are available. However, recent studies question the feasibility of predictions far beyond the location of training data. Innovation: We propose a novel workflow for spatial predictive mapping that lever-ages  recent  developments  in  this  field  and  combines  them  in  innovative  ways  with  the aim of improved model transferability and performance assessment. We demonstrate,  evaluate  and  discuss  the  workflow  with  data  from  recently  published  global  environmental maps. Main conclusions: Reducing predictors to those relevant for spatial prediction leads to  an  increase  of  model  transferability  and  map  accuracy  without  a  decrease  of  prediction  quality  in  areas  with  high  sampling  density.  Still,  reliable  gap-free  global  predictions were not possible, highlighting that global maps and their evaluation are hampered by limited availability of reference data.

Details zur Publikation

Veröffentlichungsjahr: 2023
Sprache, in der die Publikation verfasst ist: Englisch
Link zum Volltext: https://onlinelibrary.wiley.com/doi/full/10.1111/geb.13635