wiki:Documentation/Ancillary

Development Activities

Documentation

Source Code

Reference Simulations

Group Activities

Ancillary Data

  1. Ancillary Data
    1. 1. Vegetation information
      1. 1.1 Olson map
      2. 1.2 PFT maps
    2. 2. Soil texture and other soil properties
    3. 3. Irrigation and Floodplains
    4. 4. Slope
    5. 5. For routing
    6. CF-conformant files

This page describes the Ancillary data needed to describe the continental surfaces in ORCHIDEE. All the files are expected to be in a CF-compliant NetCDF format and some guidelines for producing these files are given at the end.

The most common forcing files are stored in the shared accounts in IGCM/SRF directory. The shared accounts are found:

  • LSCE, obelix : /home/orchideeshare/igcmg/IGCM/SRF
  • IPSL Ciclad : /projsu/igcmg/IGCM/igcmg/IGCM/SRF
  • At TGCC : IGCM/SRF found in igcmg shared space in work
  • At IDRIS : IGCM/SRF found in rech/psl/commun/IGCM on the workdir

If you don't find the folder at TGCC or IDRIS, ask your collegues or send an email to orchidee-help (we're not allowed to put the complete adress on a public web page)

1. Vegetation information

ORCHIDEE model can read vegetation map based on Olson categories or on PFT categories. When a PFT map is used as an input, the parameter land_use in the .def file should be set to TRUE, else to FALSE.

1.1 Olson map

A map at 5km resolution on a Goode homolosine projection with the dominant Olson class has been generated by Nicolas Viovy based on a 1km IGBP map. The map is available on the common repository of ORCHIDEE forcing files on CCRT, IDRIS and other platforms. It is also available ​here. This map contains 94 land categories. The conversion from Olson to PFT categories is done within the code of ORCHIDEE. Up to the version 1.9.5.2, there was a bug when converting Olson category 79 (warm C4 woody savanna). It was distributed into tropical broad-leaved raingreen (40%) and C3 grass (40%) while it should have been tropical broad-leaved raingreen (40%) and C4 grass (40%). This bug has been fixed in the version 1.9.6.

1.2 PFT maps

1.2.1 Based on Olson

Static maps

A Python script have been developed by Martial Mancip in order to convert 'offline' Olson-based map to PFT-based map. For this purpose, an Olson map has been generated at a 0.5 deg resolution, in which fractions of Olson categories are defined for each cell (and not only the dominant category). The 0.5° Olson map is available ​here. The PFT map generated by this script is available ​????.

Dynamic maps

The PFT map based on Olson categories has been combined with some historical trends and future projections datasets of Land-Use changes. This work has been carried out by Nathalie de Noblet within the frame of the AR5 IPCC exercise. The historical dataset is the HYDE 3.0 dataset (Klein Goldewijk) and the future projections are delivered by the four Integrated assessments models for the four Representative Concentration Pathways (RCP): MESSAGE (RCP 8.5), AIM (RCP 6), MiniCAM (RCP 4.5), IMAGE (RCP 3.0). A report and a presentation on the methodology developped have been written by Nathalie. Both documents are available [

1.2.2 Maps based on recent Satellite Land Cover products

ESA CCI Land Cover dataset (​https://www.esa-landcover-cci.org/) is used to produce PFT maps for the ORCHIDEE model. Transformation of initial data is done in the following steps:

  1. Aggregation ESA CCI LC (300 m) → Generic PFTs (given resolution), using:
    • aggregation tool "lc-user-tools" (v4.3, release 03/12/2018)
    • yearly ESA CCI land cover map v2.0.7b (1992-2015) or v2.1.1 (2016-2019) at 300 m resolution with 38 land cover classes
    • static Köppen-Geiger climate zone map at 300 m resolution grouped into 5 climate zones
    • cross-walking table (v2.4) matching ESA CCI classes with generic PFTs
    • additional cross-walking table (v2.4) refining the conversion of several ESA CCI classes depending on climate zone
  1. Aggregation Generic PFTs (given resolution) → ORCHIDEE PFTs (given resolution), using:
    • python script "gen2orc.py": link
    • yearly Generic PFT maps (1992-2019)
    • static Köppen-Geiger climate zone map (Rubel et al., 2017) for splitting trees
    • static C3/C4-partitioning map (C. Still et al., 2018) for splitting grasses
    • yearly land-use maps (LUH2-GCB2022) for splitting crops
    • static water bodies map (ESA CCI WB v4.0) for splitting inland/coastal waters and defining land-sea mask
  1. Merge ORCHIDEE PFTs (2010, 0.25°) + LUH2 (850-2022, 0.25°), using:
    • python script "mergeLUH.py": link
    • single ORCHIDEE PFT map for 2010 (0.25°)
    • yearly land-use dataset (850-2022, 0.25°, LUH2-GCB2022)
    • static Köppen-Geiger climate zone map (Rubel et al., 2017) for splitting trees
    • static C3/C4-partitioning map (C. Still et al., 2018) for splitting grasses
    • static water bodies map (ESA CCI WB v4.0) for defining land-sea mask

The full description, comparison plots and production scripts are available on the devoted web-page: ​https://orchidas.lsce.ipsl.fr/dev/lccci

1.2.3 Global Land Use modeling data formatted for ORCHIDEE

1.2.4 Future projections based on RCP's

2. Soil texture and other soil properties

Two soil texture maps are available on the shared accounts and compatible with all versions of ORCHIDEE:

  • soils_param.nc: Zobler map at 1° resolution, with 7 soil texture classes reduced in ORCHIDEE into three texture classes (coarse, medium, fine)
  • soils_param_usda.nc: Reynolds map, at 5 arc-min resolution (1/12°), with 12 USDA texture classes directly usable in ORCHIDEE

More details in ​http://forge.ipsl.fr/orchidee/raw-attachment/wiki/Documentation/UserGuide/eqs_hydrol.pdf, section "6. Soil Texture".

In branch 2.2 (starting r6954), we've introduced the possibility to describe the soil properties with other maps, which involves the keyword SPMIPEXP:

  • alternative soil texture maps (default value of SPMIPEXP, so you only have to declare the files in sechiba.card):
    • soils_param_SoilGrids_5km.nc and soils_param_SoilGrids_halfdeg.nc: the 1km SoilGrids? maps of the percent sand,silt,clay have been converted to the 12 USDA classes, and upscaled at two resolutions :
    • soils_param_usda_with_clay_oxisol.nc: modified Reynolds soil texture map, in which the clay texture (nsc=12) is split in two classes, the swelling Clays which keep the class index 12 and the corersponding soil parameters, and the non-swelling Clays (Oxisols), corresponding to a new soil class 13, with new soil hard-coded parameters, (cf the ​PhD thesis of Salma Tafasca, 2020, chapter 4).
  • maps of soil parameters (Ks, porosity, residual moisture content, Van Genuchten n and alpha), using KEYWORD SPMIPEXP = maps; you only need to declare the selected files (among the list below) in your sechiba.card
    • hydrol_params_spmip_halfdeg.nc: this file contains soil hydraulic parameters given by the SP-MIP team
    • hydrol_params_Zhang2018_halfdeg.nc: ks, theta_s and theta_r are taken from Zhang et al. (2018), n and alpha are obtained using the SoilGrids_halfdeg.nc texture map and the Schaap et al. (2001) pedo-transfer function. NA is replaced with the parameters of the Loam class. More details in the PhD thesis of Salma Tafasca chapter 5, where this map is used in the simulation SIMUSOC2
    • hydrol_params_with_porosity_Reynolds.nc: contains soil hydraulic parameters derived from the Reynolds soil texture map, using the pedo-transfer function from Carsel & Parrish, except for porosity, which is taken from the porosity map of Reynolds et al. (2000). Field capacity and wilting point are also calculated based on this porosity. More details in the PhD thesis of Salma Tafasca chapter 5, where this map is used in the simulation SIMUSOC1.
    • hydrol_params_salma_SOC3.nc: contains soil hydraulic parameters derived from soil texture (Reynolds + Carsel & Parrish 1988) and soil organic content (from an ORCHIDEE simulation described in Qiu et al., 2019). The construction of these maps is decribed in the PhD thesis of Salma Tafasca chapter 5, where this map is used in the simulation SIMUSOC3.
  • you can finally impose a uniform texture over the globe owing to two keywords: SPMIPEXP = unif, then UNIF_CASE = a to d ('a' for loamy sand, 'b' for loam, 'c' for silt, 'd' for swelling clay)

More details in SRF/SOIL/README_SPMIP

3. Irrigation and Floodplains

Several files exist with overlapping information:

4. Slope

Slope is required as input to hydrol.f90 to constrain the reinfiltration of surface runoff. The available map is cartepente2d_15min.nc, at 15 arc-min resolution (1/4°). Most likely, it is based on the ETOPO5 or ETOPO2 DEM from the USGS. More details in ​http://forge.ipsl.fr/orchidee/raw-attachment/wiki/Documentation/UserGuide/eqs_hydrol.pdf, section "4.2.3 Subgrid scale".

5. For routing

The routing module requires several information at a higher resolution than ORCHIDEE:

  • the drainage directions between grid-cells (trip)
  • the topographic or slope index (topoind) based on the local slope and length of the routing grid-cells
  • for each grid-cell, the ID of the large river basin it belongs to (basins)

More info on ​https://forge.ipsl.fr/orchidee/attachment/wiki/GroupActivities/Training/1411_ORCHIDEE_training_course_routing-expanded.2.pdf

Two datasets are available on the shared account:

CF-conformant files

Here are some recommendations to follow about the creation of forcing files. The idea is to achieve a kind of CF-like standardization of the files. Documentation/Forcings/CfStandard

Last modified 2 months ago Last modified on 2024-10-15T14:49:18+02:00

Attachments (2)

Download all attachments as: .zip