Radiosonde humidity bias correction over the West African region for the special AMMA reanalysis at ECMWF

Abstract

During the African Monsoon Multidisciplinary Analysis (AMMA) field experiment in 2006 there was a large increase in the number of radiosonde data over West Africa. This has the potential of improving the numerical weather prediction (NWP) analysis/forecast and the water budget studies over that region. However, it is well known that the humidity from radiosondes can have some errors depending on sonde type, relative humidity (RH), temperature and the age of the sensor and can give rise to dry biases that are typically between 5% and 30% for RH. Three main sonde types were used in the AMMA field experiment: Vaisala RS80A, Vaisala RS92 and MODEM. In this article, a new empirical method is presented by using the operational European Centre for Medium-Range Weather Forecasts (ECMWF) short-range forecast as an intermediary dataset for computing biases. The validation of the correction method using global positioning system (GPS) total columnar water vapour (TCWV) confirms that the method is able to correct for a large part of the dry biases associated with the different sonde types. Results from analysis experiments show how the correction of humidity is particularly important in the West African region due to its impact on the development of convection in NWP models. The proposed radiosonde humidity bias correction has been applied to the special AMMA reanalysis experiment performed at ECMWF for the 2006 West African wet monsoon season. This is expected to benefit a wide number of AMMA-related studies that make use of the reanalysis, in particular those focusing on the water cycle.

Assimilation of water‐vapour airborne lidar observations: impact study on the COPS precipitation forecasts

“…They were used to compute the bias and RMSE of the three assimilation cycles over the whole period. Even though the additional COPS soundings provide the most frequent and accurate moisture observations available, they also might have errors (Agusti-Panareda et al, 2009;Nuret et al, 2008;Bock and Nuret, 2009). However, the different comparisons carried out during COPS between RS, dropsondes, groundbased and airborne lidar data did not reveal any suspicious bias in either the RS data or the lidar data.…”

A process-based evaluation of dust-emitting winds in the CMIP5 simulation of HadGEM2-ES

“…The skill of ERA-Interim in NLLJ forecasting is tested with a contingency table using quality controlled radiosondes launched during the African Monsoon Multidisciplinary Analysis (AMMA, Redelsperger et al 2006;Parker et al 2008;Agust-Panareda et al 2009). The longest record of radiosondes around 00 UTC is available at Agadez, Niger (16 • N, 7 • E) for January-October, which is located in an area of frequent NLLJ formation.…”

Unified treatment of dry convective and stratocumulus‐topped boundary layers in the ECMWF model

“…Here we extend the evaluation to a realistic diurnal cycle in the full forecast system compared to profile observations. The African Monsoon Multidisciplinary Analysis (AMMA; Redelsperger et al, 2006) project was chosen because of the availability of high-resolution high-quality bias-corrected radiosonde data (Vaisala 92) over areas and periods of pure dry convection (Parker et al, 2008;Agustí-Panareda et al, 2009). The observation site at Niamey before the onset of the West African monsoon period in May 2006 provides such dry conditions.…”

“…The simulation is a little more wellmixed than the radiosonde data, the inversion is realistically represented, and the moisture observations near the surface are very possibly biased dry due to the location and sounding biases (Agustí-Panareda et al, 2009). The total moisture flux (Figure 6(f)) indicates that the moisture deposited above the BL comes from surface evaporation (K-flux) as well as largescale moisture convergence (K-and M-flux).…”