Cycle 37r3 Detail

Implementation Date: 15 November 2011

The new cycle includes a collection of improvements to the forecast model and the data assimilation system. The cycle also contains hourly post-processing of model data to 90 hours in support of the Boundary Conditions (BC) Optional Programme, including hourly wave data for all four forecast cycles. Hourly post-processed products will also be introduced for the European limited area wave model. Technical changes affecting three products (10m wind gusts, maximum and minimum 2m temperature) that require action by users are detailed in Annex B of this message. The main differences between cycle 37r3 and the current operational version (cycle 37r2) in terms of their meteorological impact are

Meteorological Changes

  • Modification of the entrainment/detrainment of convection
  • Modification of the supersaturation and deposition rate for clouds
  • Modification of the surface roughness
  • Bias correction of aircraft temperature observations
  • Cycling of stratospheric model error (for the weak-constraint 4D-Var)
  • Assimilation of accumulated rainfall from NEXRAD radar data from the United States
  • Assimilation of ozone observations from infrared radiances
  • Use of the latest version of the NWP-SAF radiative transfer model (RTTOV-10 including FASTEM-4)
  • Retuning of cloud detection for the advanced infrared sounder data

Changes specifically to the EPS

  • Use of the NEMO ocean model (instead of HOPE) in EPS and use of the NEMOVAR ocean data assimilation system
  • Coupling of the EPS to the ocean model also at 12 UTC as of day-10 (previously done at 00 UTC, only)

Meteorological impact

The impact of the new cycle on the performance of the forecasting system has been tested in research mode during the period 1 January to 10 May 2011, and in pre-operational runs since 11 May 2011. The new cycle shows benefit in terms of objective upper-air scores in the early forecast range in both hemispheres: the geopotential scores are improved throughout the troposphere, with some degradation at the 100 hPa level. Tropical wind scores in the lower troposphere are significantly improved, specifically leading to reduced systematic errors in the strength of the low level inflow to the African and Indian monsoons. The change in surface roughness has led to reduced 10m wind speed over land and reduced diurnal amplitude in T2m, both of which result in generally reduced bias against observations. Cold biases in T2m have been reduced in specific winter-time situations where the amount of super-cooled cloud-water has been increased. The main impact on the EPS is improved probabilistic scores for the tropical winds throughout the forecast range.

Newsletter No.130 Winter2011/12: news items on pages 4 and 13-14

See our Related News Item (24 Nov 2011)


An updated version of the ECMWF forecasting system (known as cycle 37R3) was implemented on 15 November 2011, containing a collection of improvements to the forecast model, the data assimilation and the ensemble prediction system.

The main changes to the atmospheric model are the modifications of entrainment and detrainment in convection, of supersaturation and deposition rates in clouds, and of surface roughness. The changes to the convection scheme produce a significant positive impact on near-surface temperature and wind forecasts in the Tropics

The update of supersaturation and deposition rates in clouds mostly affects the amount of liquid water in supercooled clouds, leading to reduced night-time cooling and thus reduced cold temperature biases in certain winter situations at high latitudes.

Variable bias corrections for aircraft temperature observations, the assimilation of rainfall measurements from NEXRAD radars over the United States, and the assimilation of ozone measurements from infrared satellite instruments were also introduced. These changes increase the number of observations used in the data assimilation system and enhance the physical consistency of temperature and moisture analysis.
The main changes to the ensemble prediction system (EPS) are the use of the NEMO ocean model (instead of HOPE) and the NEMOVAR ocean data assimilation system. Now, the 12 UTC EPS run is also coupled to the ocean model from day10 onwards; previously only the 00 UTC run was coupled.

Users of ECMWF’s web pages may have noticed that most of the displayed forecast charts are now being produced with the new version of the ECMWF visualisation software, Magics++.