A gridded dataset of belowground autotrophic respiration from 1980 to 2012 in global terrestrial ecosystems upscaling of observations

This data repository contains (1) yearly global autotrophic respiration (RA) dataset from 1980 to 2012 with a spatial resolution of 0.5°; (2) original field observations to develop Random Forest (RF) model; (3) main R codes to produce RA database.

Model description:
The globally gridded RA database was developed by Random Forest (RF) with 449 field observations (see “dataset.csv” in this repository, updated from Bond-Lamberty and Thomson, 2018) using 11 global variables, including gridded temperature, precipitation, diurnal temperature range, potential evapotranspiration, Palmer Drought Severity Index, nitrogen deposition, downward shortwave radiation, soil carbon content, soil nitrogen density, soil water content, land cover.

Dataset information:
Dataset name: “Respiration_autotrophic_belowgroud_glob_1980_2012_yr_half_dgree.nc”
Which means globally belowground autotrophic respiration from 1980 to 2012 with a spatial resolution of 0.5° at a yearly step.
Units: g C m-2 yr-1
Format: network Common Data Form (netCDF)
Spatial coverage: 90S-90N, 180W-180E
The “dataset.csv” file is the field observation from peer review publications combining Global Soil Respiration Database (SRDB v4, Bond-Lamberty and Thomson, 2018), which is publicly available at https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=1578. Besides, The database was further updated using observations collected from the China Knowledge Resource Integrated Database (www.cnki.net) up to November 2018 according to the criteria of SRDB. This dataset is provided in format of “.csv”.

R codes:
10fold_CV_RA.txt:
10-fold CV for RA
Annual_variability_RA.txt:
annual variability for global RA
CMP_RA.txt: comparing
RF-RA and Hashimoto2015-RA using CMP approach
Ra_DD_CC_plot.txt:
plotting the comparing results from CMP
RA_MAT_MAP_anomaly.txt:
plotting and modelling the relationship between temperature/precipitation
anomalies and RA

RGB_plot.txt:
deriving RGB plot to detecting the relative importance of temperature,
precipitation and shortwave radiation.