Towards a remote-sensing-driven model of isoprene emissions from Alpine tundra

This study investigates isoprene emissions in a high-latitude Alpine tundra ecosystem, focusing on using near-field remote sensing of surface temperatures, the photochemical reflectance index (PRI) and normalized difference vegetation index (NDVI), and meteorological measurements to model these emis...

Descripción completa

Detalles Bibliográficos
Autores: Westergaard-Nielsen, A., Maigaard, R. S., Davie-Martin, C. L., Seco, Roger, Holst, T., Pirk, N., Laursen, S. N., Rinnan, R.
Tipo de recurso: artículo
Estado:Versión publicada
Fecha de publicación:2025
País:España
Institución:Consejo Superior de Investigaciones Científicas (CSIC)
Repositorio:DIGITAL.CSIC. Repositorio Institucional del CSIC
OAI Identifier:oai:digital.csic.es:10261/401810
Acceso en línea:http://hdl.handle.net/10261/401810
https://api.elsevier.com/content/abstract/scopus_id/105016992283
Access Level:acceso abierto
Palabra clave:Remote sensing
Alpine
BVOC
Environment
Meteorology
http://metadata.un.org/sdg/11
http://metadata.un.org/sdg/3
http://metadata.un.org/sdg/6
Ensure healthy lives and promote well-being for all at all ages
Ensure availability and sustainable management of water and sanitation for all
Make cities and human settlements inclusive, safe, resilient and sustainable
Descripción
Sumario:This study investigates isoprene emissions in a high-latitude Alpine tundra ecosystem, focusing on using near-field remote sensing of surface temperatures, the photochemical reflectance index (PRI) and normalized difference vegetation index (NDVI), and meteorological measurements to model these emissions. Isoprene is a key biogenic volatile organic compound (BVOC) emitted by select plants, which can impact atmospheric chemistry and climate. Increased temperatures, particularly in high latitudes, may enhance isoprene emissions due to extended growing seasons and heightened plant stress. The research was conducted in Finse, Norway, where isoprene and CO<inf>2</inf> fluxes were measured with eddy covariance alongside spectral and meteorological data, and surface temperature. A random forest (RF) model was developed to predict isoprene fluxes, considering the variable importance of different environmental factors. The results showed that surface temperature and CO<inf>2</inf> flux were consistently important predictors, across three differential temporal data aggregations (hourly, daily, weekly), while the PRI demonstrated low predictive power, possibly due to the heterogeneous vegetation and variable light conditions. The NDVI was more effective than anticipated, likely linked to phenological changes in vegetation. Model performance varied with temporal resolution, with weekly data achieving the highest predictive accuracy (R<sup>2</sup> up to 0.76). The RF model accurately reflected seasonal emission patterns but underestimated short-term peaks, suggesting the potential to combine machine learning with process-based modelling. This research highlights the promise of proxy data from remote sensing for scaling BVOC emission models to regional levels, essential for understanding climate impacts in Arctic ecosystems.