Data and code for "Large-scale remote sensing analysis reveals an increasing coupling of grassland vitality to atmospheric water demand"

Data and code for 
"Large-scale remote sensing analysis reveals an increasing coupling of grassland vitality to atmospheric water demand"

All R code used for the analysis is provided in the folder code. 
Data and intermediate results are provided or stored in the folders data and tmp_data.
All results including figures will be stored in the folder results. 

R version: 4.3.1

To carry out the entire analysis the code should be run in the provided order:

1) Code to run non-metric multidimensional scaling (NMDS) for habitat groups and 
produce Fig. 1b (habitat map and legend for Fig 1a: data/eunis_gl_habitat_ger_990m.tif,eunis_gl_habitat_ger_990m_legend.clr)
 
2) Code to generate grassland vitality maps and time series from 1985 to 2021 (Fig. 3). 
Grassland vitality maps on 30m for all grasslands in Germany provided in data/glv_1985-2021.zip.

3) Code to model relation of grassland vitality to five drought indices (VPD, temperature, CWB, soil moisture, precipitation),
output are Fig. 4, Fig. S1, Tab. 1.

4) Code for trend analysis of drought sensitivity based on 5-, 10-, and 15-year moving windows, output are Fig. 5, Fig. S2. 

5) Code to model drought sensitivity of grassland habitat groups and habitat types, output are Fig. 6 and table with sensitivity per habitat type.