Navigation path


EXPERIMENTSRESULTSMODELSPARTICIPANTS Up Print version Decrease text Increase text Home

Active submission period: May 2008 - May 2009

On-line results publication: June 2011

Scientific Publication:

RAMI4PILPS Publication J.L. Widlowski, B. Pinty, M. Clerici, Y. Dai, M. De Kauwe, K. de Ridder, A. Kallel, H. Kobayashi, T. Lavergne, W. Ni-Meister, A. Olchev, T. Quaife, S. Wang, W. Yang, Y. Yang, and H. Yuan (2011), RAMI4PILPS: An intercomparison of formulations for the partitioning of solar radiation in land surface models, Journal of Geophysical Research, 116, G02019, 25, DOI: 10.1029/2010JG001511.

Bibliography: LaTeX, Word


The active submission phase of RAMI4PILPS is now closed. The purpose of the RAMI4PILPS experiment suite was to evaluate and compare the different approaches by which land surface schemes (LSS) in climate and weather prediction models quantify the radiation transfer within and beneath vegetation canopies. Given the availability of remotely-sensed, multi-annual, global datasets of radiative surface fluxes, that allow to capture both the spatial and temporal evolution of the energy partitioning (albedo, absorption, transmission) of terrestrial surfaces, RAMI4PILPS can be envisaged as a quality control mechanism that assesses the appropriateness of radiative flux formulations in current LSS in the light of future assimilation efforts of remote sensing products into climate and weather prediction models. Benefits to PILPS land surface models and their users include: 1) to quantify the typical errors associated with different modes of estimating the radiative surface fluxes in LSS; 2) to identify the impact that structural and spectral sub-grid variability may have on these flux estimates; and 3) to verify the conservation of energy at the level of the surface, as well as inconsistencies arising from the derivation of flux quantities from different sources with different levels of assumptions/simplifications. RAMI4PILPS will assess the quality of the prescribed/simulated radiative fluxes in LSS by direct comparison with reference solutions obtained from credible models identified during the third phase of the RAMI benchmarking exercise.

grasslandsclosed forest canopiesshrublandsopen forest canopies
closed/dense forest canopies
open/sparse forest canopies
Given detailed 1-D canopy descriptions,
what are the values of the three fluxes R, A, and T?
Given detailed 3-D canopy descriptions and R
how is the remaining energy split between A and T?

Participation in RAMI4PILPS is open to all interested parties (please read our disclaimer). RAMI4PILPS focuses entirely on radiative surface fluxes (i.e., canopy absorption, transmittance and reflectance) in the shortwave domain (more specifically, solar radiation in the visible and near-infrared spectral regions, separately). Thus, the RAMI4PILPS suite of experiments does not deal with radiative quantities pertinent to longwave radiation (whether latent and specific heat, or, energy arising from phase changes, heat transfer and surface emissions) nor does it involve any forcing terms or temporal evolution. Instead, RAMI4PILPS focuses exclusively on the partitioning of shortwave radiation - impinging with a given solar zenith angle at the top of a vegetation canopy - into a reflected (R), absorbed (A) and transmitted (T) flux ratio component, such that the energy balance can be written as: A = 1 - R - (1-alpha)T, where alpha is the soil albedo and T includes all orders of scattered radiation that reach the background. Two different categories of experiments have been designed: 1) structurally homogeneous (1-D) environments where participants are required to deliver all three radiative surface fluxes, R, A and T on the basis of detailed canopy descriptions, and 2) structurally heterogeneous (3-D) environments, where participants are provided with a detailed canopy description and the surface reflectance, R - often available in real application from remote sensing observations - and are requested to deliver their model's estimate of the partitioning of the remaining energy into A and T. The soil albedo is provided for all test cases in categories 1 and 2.

1-D canopies
  • grasslands: can be represented as 1-D canopies where the short vegetation covers all of the underlying (flat) background.
  • closed forest canopies: can be represented as 1-D canopies where the tall vegetation covers all of the underlying (flat) background.
3-D canopies
  • shrublands: can be represented as 3-D canopies where the vegetation is clumped into small spherical volumes close to and covering only a fraction of the underlying (flat) background.
  • open forest canopies: can be represented as 3-D canopies where the vegetation is clumped into large spherical volumes of variable height hovering above and covering only a fraction of the underlying (flat) background.

Interested participants are encouraged to inform the RAMI4PILPS coordinators and to submit their flux estimates of the above test cases using the FTP procedure outlined here. We encourage the submission of simulation/look-up-table results from stand-alone radiation transfer models as well as modules that are part of larger simulation models, be they SVAT, NWP or regional/global circulation models. Furthermore, we want to make it clear that participation is encouraged irrespective of whether your model estimates one or more of the required flux quantities (just provide those that you have access to, or, can generate: for the 3-D scenes it may actually be that some models can predict the reflectance already on the basis of the provided structural and spectral canopy properties), and also irrespective of whether your model/module does retrieve/prescribe these quantities through the usage of look-up-tables, through parameterisations, or, through actual radiation transfer calculations. In case of difficulties or uncertainties in the implementation of the prescribed test cases, please do not hesitate in contacting us via the above email. Access to the various experiment description pages can be obtained by clicking on the corresponding images and links above. In addition, a synoptic overview table is provided below to give direct access to all experiments proposed for RAMI4PILPS (left column) as well as their associated measurements (right column).

Given that RAMI4PILPS will be run like an ordinary RAMI phase, first time users may read the FAQ section to get started, as well as, the file formatting and naming convention. Prior to the submission of results, RAMI4PILPS participants are strongly encouraged to test the compliance of their model-generated files using the RAMI Online Format Checker. To do so follow the pulsating on-line format checker link in the top (red) navigation bar. Once your files are correct please follow the upload instructions in the FAQ: How do I submit my results?

Right navigation