Introduction

The scripts in this repository are a result of the research into enhanched dendroprovenancing using gridded environmental variables. Scripts are expected to be ran in sequence in ascending order. For further documentation we refer to the inline comments.

Python package requirements

astropy
fiona
gdal
joblib
matplotlib
netCDF4
numpy
pandas
rasterio
scipy
shapely
sklearn
sklearn_quantile
statsmodels
tqdm

System requirements

• >16gb RAM
• ~50gb storage space (depending on the number of chronologies)

Folder Structure

Prepare the following folders in the same directory as the python scripts:

├── 0_1a_trw_download
├── 0_1b_trw_raw
├── 0_2_CRU_TS
├── 0_3_worldclim
├── 0_4_soilgrids
│   ├── source
│   ├── prepared
├── 0_5_species_distribution
├── 1_1_prepared_trw
├── 1_2_detrended_trw
├── 1_3_chronologies
├── 2_1_random_forests
├── 2_2_modelled_chronologies
├── 3_1_year_known_out
├── 3_2_both_unknown_out
├── 4_0_general_outputs
├── temp

Data requirements

ITRDB:

Download raw TRW data for the tree species of interest from ITRDB (https://www.ncei.noaa.gov/access/paleo-search/?dataTypeId=18). To replicate the case study, download Quercus robur data with temporal overlap for the period 1901-2022. Optional: place raw download .zip in 0_1a_trw_download. Unzip and place all files in measurements folder(s) into 0_1b_trw_raw.

CRU TS:

Download CRU TS data (https://crudata.uea.ac.uk/cru/data/hrg/cru_ts_4.06/cruts.2205201912.v4.06/), and place directly into 0_2_CRU_TS. The folder should contain four files, one for each variable (pre, tmn, tmp, tmx). Example filename for pre: cru_ts4.06.1901.2021.pre.dat.

Worldclim:

Download Worldclim data (https://www.worldclim.org/data/worldclim21.html), and place directly into 0_3_worldclim. The folder should contain four folders, one for each variable (prec, tavg, tmax, tmin). Example foldername for prec: wc2.1_30s_prec.

Soilgrids:

Download Soilgrids data (easiest through WCS, see https://www.isric.org/explore/soilgrids), and place into 0_4_soilgrids/prepared. The folder should contain five TIF files (resolution: 43200x21600), one for each variable (clay, nitrogen, sand, silt, soc). Example filename for clay: clay_15_30cm.

Species distribution:

Download or create species distribution data (.shp) and place directly into 0_5_species_distribution. To replicate the case study, download from https://data.mendeley.com/datasets/hr5h2hcgg4 and copy all Quercus_robur_plg_clip files.