Preprints
https://doi.org/10.5194/egusphere-2023-2559
https://doi.org/10.5194/egusphere-2023-2559
06 Dec 2023
 | 06 Dec 2023
Status: this preprint is open for discussion.

Explainable machine learning for modelling of net ecosystem exchange in boreal forest

Ekaterina Ezhova, Topi Laanti, Anna Lintunen, Pasi Kolari, Tuomo Nieminen, Ivan Mammarella, Keijo Heljanko, and Markku Kulmala

Abstract. There is a growing interest in applying machine learning methods to predict net ecosystem exchange (NEE) based on site information and climatic parameters. In case of successful performance, it could give an excellent opportunity for gapfilling or upscaling, i.e., extrapolation of results to times and sites for which direct measurements are unavailable. There exists already quite an extensive body of research covering different seasons, time scales, number of sites, input parameters (features), and models. We apply four machine learning models to predict NEE of boreal forest ecosystems based on climatic and site parameters. We use data sets from two stations in the Finnish boreal forest and model NEE during the peak growing season and the whole year. Using Explainable Artificial Intelligence methods, we compare the most important input parameters chosen by the models. In addition, we analyze the dependencies of NEE on input parameters against existing theoretical understanding on NEE drivers. We show that even though the statistical scores of some models can be very good, the results should be treated with caution especially when applied to upscaling. In the model setup with several interdependent parameters ubiquitous in atmospheric measurements, some models display strong opposite dependencies on these parameters. This behavior might have adverse consequences if models are applied to the data sets in future climate conditions. Our results highlight the importance of Explainable Artificial Intelligence methods for interpreting outcomes from machine learning models, in particular, when a large set of interdependent variables is used as a model input.

Ekaterina Ezhova, Topi Laanti, Anna Lintunen, Pasi Kolari, Tuomo Nieminen, Ivan Mammarella, Keijo Heljanko, and Markku Kulmala

Status: open (until 15 Mar 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • RC1: 'Comment on egusphere-2023-2559', Anonymous Referee #1, 29 Jan 2024 reply
Ekaterina Ezhova, Topi Laanti, Anna Lintunen, Pasi Kolari, Tuomo Nieminen, Ivan Mammarella, Keijo Heljanko, and Markku Kulmala

Data sets

Hyytiälä, Värriö Hari et al, Kulmala et al https://smear.avaa.csc.fi/

Ekaterina Ezhova, Topi Laanti, Anna Lintunen, Pasi Kolari, Tuomo Nieminen, Ivan Mammarella, Keijo Heljanko, and Markku Kulmala

Viewed

Total article views: 310 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
220 75 15 310 9 10
  • HTML: 220
  • PDF: 75
  • XML: 15
  • Total: 310
  • BibTeX: 9
  • EndNote: 10
Views and downloads (calculated since 06 Dec 2023)
Cumulative views and downloads (calculated since 06 Dec 2023)

Viewed (geographical distribution)

Total article views: 298 (including HTML, PDF, and XML) Thereof 298 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 26 Feb 2024
Download
Short summary
ML models are gaining popularity in biogeosciences. They are applied as gapfilling methods and used to upscale carbon fluxes to larger areas based on local measurements. In this study, we use Explainable ML methods to elucidate performance of machine learning models for carbon dioxide fluxes in boreal forest. We show that statistically equal models treat input variables differently. Explainable ML can help scientists to make informed solutions when applying ML models in their research.