Experts discuss role of drag processes in NWP and climate models

Irina Sandu, Ayrton Zadra (Environment Canada)


The workshop on ‘Drag processes and their links to the large-scale circulation’, organised jointly by ECMWF, the World Climate Research Programme (WCRP) and the World Weather Research Programme (WWRP), was held at ECMWF from 12 to 15 September. Despite their importance for the large-scale circulation, to date the representation of drag processes remains a major source of uncertainty in global models. ‘Drag’ refers to the effects of friction on atmospheric flow caused by elements of the land surface, ocean waves, orography and the breaking of mountain-induced gravity waves.

The workshop aimed to assess the current state of our understanding of drag processes and their impact on the large-scale circulation on timescales from synoptic to seasonal and climate timescales. The workshop also aimed to review how these processes are represented in global models; discuss and sharpen the research challenges to be overcome in order to achieve substantial advances in this area; foster collaborations; and stimulate further research. The idea of organising this workshop partially stemmed from the WGNE (WMO Working Group on Numerical Experimentation) ‘Drag project’, which demonstrated that the main NWP and climate models differ significantly both in the representation of total surface stress (or friction), particularly in regions with orography, and in the partitioning of surface stress among various physical processes.

The workshop attracted about 50 participants from the main numerical weather prediction (NWP) and climate centres in Canada, France, Germany, Japan, the Netherlands, the UK and the US as well as from several universities. The participants included well-established scientists and early-career scientists, six of whom were partially supported by WMO.


A broad range of scientific questions were discussed through invited talks, a poster session and working group discussions. Three main themes were covered: (i) theoretical aspects of drag processes and impacts of uncertainty associated with drag processes in NWP and climate models, (ii) the representation of drag in global models (parametrizations, ancillary fields such as mean and subgrid orography etc.), and (iii) constraining drag processes through observations, reanalysis and fine-scale modelling. The working groups made numerous recommendations for further research in these areas. These include:

  • Consolidate knowledge regarding the impacts of drag processes on the large-scale circulation, e.g. by reproducing results in different models, and develop a more quantitative understanding of effects of drag on aspects of circulation, such as the mean state, stationary waves, synoptic systems. Understand what level of parametrization is required to reproduce given phenomena and whether there are processes that are currently not represented in global models.
  • Seek to further understand inter-model differences in surface stress, for example through the following activities: a survey regarding the ancillary files, in which all centres would provide details on corresponding databases and methods as well as samples of ancillary fields; numerical experiments aiming to better define the appropriate sub-grid scales for orographic fields as a function of the model's (effective) resolution; extending the WGNE Drag project by comparing the tendencies given by the various parametrizations in regions of maximum uncertainty, and by using relevant single-column model experiments.
  • Explore the use of high-resolution simulations, which can now be performed at resolutions of a few hundred metres over large regions, to help understand the underlying processes contributing to orographic drag and to constrain the parametrizations. As surface drag cannot be observed on large scales, this type of simulation could provide a reference estimate of surface drag that would be extremely valuable for improving the parametrizations used in global models.
  • Explore new methods to identify the parametrizations responsible for model errors and devise ways of optimising poorly constrained parameters that go beyond empirical tuning. These can include initial tendency diagnostics, nudging techniques, data assimilation methods, but also a more process-level-based evaluation of the phenomena represented by the parametrizations (e.g. waves vs turbulence) or the evaluation of theoretically understood far-field responses to changes in drag.
  • Make more extensive use of existing direct or indirect observations to evaluate the representation of drag processes in models. Here examples include emerging observations of momentum fluxes, gathered either in observational campaigns or at permanent supersites, and scatterometer wind data or bulk measures of drag impacts on the circulation, such as the change in wind direction throughout the boundary layer.

Workshop participants.
Workshop participants. The workshop was organised by Irina Sandu (ECMWF), Felix Pithan (AWI), Julio Bacmeister (NCAR), Andreas Dörnbrack (DLR), Ted Shepherd (University of Reading), Gunilla Svensson (MISU) and Ayrton Zadra (Environment Canada). It attracted about 50 participants from NWP and climate centres as well as universities across the world.

The workshop presentations are available on ECMWF’s website at