1 USING SCINTILLOMETERS AND CEILOMETERS FOR VALIDATION OF THE WRF- MESOSCALE MODEL G.J. Steeneveld*, O.K. Hartogensis*, A.F. Moene*, H. Klein Baltink + , and A.A.M. Holtslag* *Wageningen Univ., Wageningen, The Netherlands. + Royal Netherlands Meteorological Institute, De Bilt, The Netherlands 1. INTRODUCTION 1 & BACKGROUND Forecasting of near surface weather, species transport and dispersion, and the inversion of greenhouse gas transport on the mesoscale re- lies on the performance of the atmospheric boundary layer (ABL) and land surface scheme in limited area models (e.g. Denning et al., 2008: Gerbig et al., 2008). However, the PBL descrip- tion in NWP models still has difficulties (Steene- veld et al., 2008), especially in the stable ABL (SBL). Nighttime mixing is often overestimated and the low level jet misrepresented. During day- time the representation of ABL entrainment could be improved. All together this results in er- rors in the diurnal cycle of wind speed, direction and the thermodynamic variables (Olivié, et al., 2004; Svensson and Holtslag, 2007; Teixeira et al., 2008). Hence there is need to compare mesoscale model results with observations to understand the model limitations as well as their strengths. In the study described in this paper, the PBL schemes implemented in mesoscale model WRF are evaluated against a network of in situ obser- vations in The Netherlands. Previous studies also evaluated WRF, but these were mostly fo- cused on complex terrain, the synoptic scale (Cheng and Steenburgh, 2005) or air quality (Tie, 2007). Usually atmospheric mesoscale models are evaluated against point measurements. How- ever, then representation errors occur. Surface fluxes are calculated on a grid scale, so they also should be evaluated against observed area averaged fluxes. The innovative aspect of this study is the use of a network of scintillometers and ceilometers for model evaluation. We will compare observed surface fluxes of momentum (u * ), sensible heat (H) and evapotranspiration (L v E ), and next also the profiles of wind speed (U), potential temperature (θ), and specific hu- midity (q). The second aim is to compare mod- elled diurnal cycle between the MRF scheme (Troen and Mahrt, 1986) and its improved equivalent YSU (Noh et al., 2003). 2. OBSERVATIONS A scintillometer is an instrument that consists of a light transmitter and a receiver. The instrument records the integrated effect of the turbulent per- turbations of the air’s refractive index (n), and its 1 G.J. Steeneveld, Wageningen Univ., Meteorology and Air Quality Group, P.O. box 47, 6700 AA Wageningen, The Netherlands. E-mail: Gert-Jan.Steeneveld@wur.nl structure parameters (C n 2 ). Monin-Obukhov the- ory is used to convert C n 2 to area-averaged sur- face fluxes of heat, using 10 m wind as input. We use a optical Large Aperture Scintillometers (LAS) which operate on a scale of ~500-5000m in regions of the Netherlands with different vege- tation types. See Meijninger et al. (2002) for more information on the LAS. A ceilometer is an instrument that measures the ABL height (h) using laser or other light techniques. In addition to these innovative in- struments, we also evaluate the model against Cabauw tower observations (e.g. Beljaars and Bosveld, 1997), and routine micrometeorological observations. 3. MODEL SETUP & CASE DESCRIPTION We have selected two cloud free days: 11 June 2006 with strong winds (~4 ms -1 at 10m), and 30 June-2 July 2006, which is the GABLS3 episode. The area consists of mainly grassland and is flat and relatively homogeneous. Also, the area has a large water supply and thus a high soil mois- ture availability. For these simulations, the initial and boundary conditions (every 6 h) were pro- vided by NCAR-FNL. However, using ECMWF as boundary conditions provided similar results. WRFv3 was run in an area of 1000 x 1000 km with a grid size of 16 km. In this domain, we nested 1 domain with a grid spacing of 4, km to minimize model errors due to lack of horizontal resolution. Moreover, the U.S. Geological Survey provided the land surface properties for WRF such as soil moisture availability, surface rough- ness, and land use. WRF was run with 3 different ABL schemes. First, we use the so-called MRF scheme (Troen and Mahrt, 1986; Hong and Pan, 1996) which utilizes a prescribed cubic eddy diffusivity profile with height, with the magnitude depending on the characteristic velocity scale at the surface layer. This scheme allows for non-local heat transport during the day. This extension is needed to represent transport by large eddies on the scale of the ABL itself, instead of local trans- port. A well-known drawback of this widely used scheme is excessive daytime ABL top entrain- ment, and overestimation of the turbulent trans- port at night (e.g. Vila et al., 2002; Steeneveld et al., 2008). The 2 nd scheme is an extension MRF, (so called YSU). The extensions consist of a) inclu- sion of prescribed entrainment rate at the ABL top, b) non-local transport of momentum, and c) Prandtl number (K M /K H ) depending on height