Abstract
AbstractThis paper discusses different feature selection methods and CO2 flux data sets with a varying quality‐quantity balance for the application of a Random Forest model to predict daily CO2 fluxes at 250 m spatial resolution for the Rur catchment area in western Germany between 2010 and 2018. Measurements from eddy covariance stations of different ecosystem types, remotely sensed vegetation data from MODIS, and COSMO‐REA6 reanalysis data were used to train the model and predictions were validated by a spatial and temporal validation scheme. Results show the capabilities of a backwards feature elimination to remove irrelevant variables and an importance of high‐quality‐low‐quantity flux data set to improve predictions. However, results also show that spatial prediction is more difficult than temporal prediction by reflecting the mean value accurately though underestimating the variance of CO2 fluxes. Vegetated parts of the catchment acted as a CO2 sink during the investigation period, net capturing about 237 g C m−2 y−1. Croplands, coniferous forests, deciduous forests and grasslands were all sinks on average. The highest uptake was predicted to occur in late spring and early summer, while the catchment was a CO2 source in fall and winter. In conclusion, the Random Forest model predicted a narrower distribution of CO2 fluxes, though our methodological improvements look promising in order to achieve high‐resolution net ecosystem exchange data sets at the regional scale.
Highlights
Land use changes are important drivers of anthropogenic climate change
The Random Forest model predicted a narrower distribution of CO2 fluxes, though our methodological improvements look promising in order to achieve high-resolution net ecosystem exchange data sets at the regional scale
backward feature elimination (BFE) always performed better than no feature selection, indicating that BFE is more suitable than forward feature selection (FFS) or no feature selection for this analysis
Summary
Land use changes are important drivers of anthropogenic climate change. For example, deforestation or afforestation can highly affect the carbon uptake and storage capacities of an ecosystem (Schimel et al, 2001). Net ecosystem exchange (NEE), the difference between carbon dioxide (CO2) uptake through photosynthesis and respiration within an ecosystem (Luyssaert et al, 2007), represents a major feature of the global carbon cycle and, helps to assess ecosystem services and the impact of land use changes on them (negative NEE = CO2 uptake, positive NEE = CO2 emission) (Abdalla et al, 2013; Schmitt et al, 2010; Xu et al, 2017). Top-down approaches for spatial NEE assessment include global atmospheric inversion models from satellites such as GOSAT and OCO-2 Wang et al, 2019), which are especially useful for areas with limited or no EC coverage (Kondo et al, 2015) but are restricted to a coarse spatial resolution. Bottom-up approaches scaling up EC measurements are expedient to quantify CO2 fluxes for larger areas (Denman et al, 2007; Xiao et al, 2012), though they are challenging due to the high spatiotemporal variability of those fluxes (Borchard et al, 2015; Kondo et al, 2017)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
