Articles | Volume 15, issue 14
Research article
25 Jul 2018
Research article |  | 25 Jul 2018

Does predictability of fluxes vary between FLUXNET sites?

Ned Haughton, Gab Abramowitz, Martin G. De Kauwe, and Andy J. Pitman

Data sets

The FLUXNET 2015 dataset Fluxdata

The CRU TS4.01 data Centre for Environmental Data Analysis

The BioClim data WorldClim - Global Climate Data

The model and analysis code GitHub

naught101/empirical_lsm: v1.1 Zenodo

Short summary
This project explores predictability in energy, water, and carbon fluxes in the free-use Tier 1 of the FLUXNET 2015 dataset using a uniqueness metric based on comparison of locally and globally trained models. While there is broad spread in predictability between sites, we found strikingly few strong patterns. Nevertheless, these results can contribute to the standardisation of site selection for land surface model evaluation and help pinpoint regions that are ripe for further FLUXNET research.
Final-revised paper