Remote sensing is a promising tool for detecting invasive alien plant species. Mapping and monitoring those species requires accurate detection. So far, most studies relied on models that are locally calibrated and validated against available field data. Consequently, detecting invasive alien species at new study areas requires the acquisition of additional field data which can be expensive and time-consuming. Model transfer might thus provide a viable alternative. Here, we mapped the distribution of the invasive alien bryophyte Campylopus introflexus to i) assess the feasibility of spatially transferring locally calibrated models for species detection between four different heathland areas in Germany and Belgium and ii) test the potential of combining calibration data from different sites in one species distribution model (SDM). In a first step, four different SDMs were locally calibrated and validated by combining field data and airborne imaging spectroscopy data with a spatial resolution ranging from 1.8 m to 4 m and a spectral resolution of about 10 nm (244 bands). A one-class classifier, Maxent, which is based on the comparison of probability densities, was used to generate all SDMs. In a second step, each model was transferred to the three other study areas and the performance of the models for predicting C. introflexus occurrences was assessed. Finally, models combining calibration data from three study areas were built and tested on the remaining fourth site. In this step, different combinations of Maxent modelling parameters were tested. For the local models, the area under the curve for a test dataset (test AUC) was between 0.57–0.78, while the test AUC for the single transfer models ranged between 0.45–0.89. For the combined models the test AUC was between 0.54–0.9. The success of transferring models calibrated in one site to another site highly depended on the respective study site; the combined models provided higher test AUC values than the locally calibrated models for three out of four study sites. Furthermore, we also demonstrated the importance of optimizing the Maxent modelling parameters. Overall, our results indicate the potential of a combined model to map C. introflexus without the need for new calibration data

Skowronek, S.; Van De Kerchove, R.; Rombouts, B.; Aerts, R.; Ewald, M.; Warrie, J.; Schiefer, F.; Garzon-Lopez, C.; Hattab, T.; Honnay, O.; Lenoir, J.; Rocchini, D.; Schmidtlein, S.; Somers, B.; Feilhauer, H. (2018). Transferability of species distribution models for the detection of an invasive alien bryophyte using imaging spectroscopy data. INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION, 68: 61-72. doi: 10.1016/j.jag.2018.02.001 handle: http://hdl.handle.net/10449/54445

Transferability of species distribution models for the detection of an invasive alien bryophyte using imaging spectroscopy data

Rocchini, D.;
2018-01-01

Abstract

Remote sensing is a promising tool for detecting invasive alien plant species. Mapping and monitoring those species requires accurate detection. So far, most studies relied on models that are locally calibrated and validated against available field data. Consequently, detecting invasive alien species at new study areas requires the acquisition of additional field data which can be expensive and time-consuming. Model transfer might thus provide a viable alternative. Here, we mapped the distribution of the invasive alien bryophyte Campylopus introflexus to i) assess the feasibility of spatially transferring locally calibrated models for species detection between four different heathland areas in Germany and Belgium and ii) test the potential of combining calibration data from different sites in one species distribution model (SDM). In a first step, four different SDMs were locally calibrated and validated by combining field data and airborne imaging spectroscopy data with a spatial resolution ranging from 1.8 m to 4 m and a spectral resolution of about 10 nm (244 bands). A one-class classifier, Maxent, which is based on the comparison of probability densities, was used to generate all SDMs. In a second step, each model was transferred to the three other study areas and the performance of the models for predicting C. introflexus occurrences was assessed. Finally, models combining calibration data from three study areas were built and tested on the remaining fourth site. In this step, different combinations of Maxent modelling parameters were tested. For the local models, the area under the curve for a test dataset (test AUC) was between 0.57–0.78, while the test AUC for the single transfer models ranged between 0.45–0.89. For the combined models the test AUC was between 0.54–0.9. The success of transferring models calibrated in one site to another site highly depended on the respective study site; the combined models provided higher test AUC values than the locally calibrated models for three out of four study sites. Furthermore, we also demonstrated the importance of optimizing the Maxent modelling parameters. Overall, our results indicate the potential of a combined model to map C. introflexus without the need for new calibration data
Campylopus introflexus
Heath star moss
Hyperspectral
Maxent
Dune ecosystem
Model transfer
Settore BIO/03 - BOTANICA AMBIENTALE E APPLICATA
2018
Skowronek, S.; Van De Kerchove, R.; Rombouts, B.; Aerts, R.; Ewald, M.; Warrie, J.; Schiefer, F.; Garzon-Lopez, C.; Hattab, T.; Honnay, O.; Lenoir, J.; Rocchini, D.; Schmidtlein, S.; Somers, B.; Feilhauer, H. (2018). Transferability of species distribution models for the detection of an invasive alien bryophyte using imaging spectroscopy data. INTERNATIONAL JOURNAL OF APPLIED EARTH OBSERVATION AND GEOINFORMATION, 68: 61-72. doi: 10.1016/j.jag.2018.02.001 handle: http://hdl.handle.net/10449/54445
File in questo prodotto:
File Dimensione Formato  
IJAG_2018.pdf

solo utenti autorizzati

Tipologia: Versione editoriale (Publisher’s layout)
Licenza: Tutti i diritti riservati (All rights reserved)
Dimensione 1.83 MB
Formato Adobe PDF
1.83 MB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10449/54445
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 21
  • ???jsp.display-item.citation.isi??? 20
social impact