The improvement of land use and land cover representation in regional climate models

Abstract

Land use and land cover change (LULCC) is one of the essential variables in climate modelling. The accurate representation of feedback mechanisms between the land surface and the atmosphere is crucial for Regional Climate Models (RCMs) to produce reliable information about the past, present and future climate. Several LULC products are available and used in RCM simulations. However, the LULC products used often lack the required spatial and temporal resolution and, most importantly, thorough validation. In order to overcome the short-comings of exiting LULC datasets this thesis introduces a newly developed, high-resolution plant functional type (PFT) map for Europe for the use in regional climate models, LANDMATE PFT 2015 Version 1.0. The map is generated translating high-resolution LULC data (ESA-CCI LC) into PFTs for the European continent. High-resolution climate data is used to differentiate the translation of LULC classes for 30 climate zones depending on temperature and precipitation data. Under consideration of the 30 climate zones, the LULC classes from the ESA-CCI LC map 2015 are translated into PFTs using a newly developed cross-walking procedure. The validation of the map is provided through comparison of LANDMATE PFT 2015 with an extensive ground truth database for the European continent. The newly developed validation workflow makes it possible to compare fractional LULC maps to a ground truth dataset, which is provided as point samples. The validation results show improvement of LANDMATE PFT in comparison to the ESA-CCI default PFT map and provide regional quality information to regional climate modellers. The workflow is tailored to the validation of fractional LULC maps and can be applied to already existing LULCC or PFT map products in order to tackle the challenge that arises from uncertainty in LULCC input in RCMs.