1 17B.6 LONG TERM OBSERVATIONS OF LONG WAVE RADIATIVE FLUX DIVERGENCE IN THE STABLE BOUNDARY LAYER OVER LAND G.J. Steeneveld, M.J.J. Wokke, C.D. Groot-Zwaaftink, S. Pijlman, B.G. Heusinkveld, A.F.G. Jacobs, and A.A.M. Holtslag Wageningen University, Wageningen, The Netherlands. 1. INTRODUCTION 1 & BACKGROUND After sunset, the Earth’s surface cools due to long wave radiation emission to space, and a stable atmospheric boundary layer (SBL) develops. The SBL structure develops due to different physical processes: turbulent mixing, radiative cooling, the interaction with the land surface, gravity waves, katabatic flows, etc. Despite previous research ef- forts, these processes and their interactions are insufficiently understood. Consequently, the SBL is poorly represented in atmospheric models. Numerical weather forecast (NWP) and climate models experience serious errors for the SBL (e.g. Viterbo et al., 1999; Dethloff et al., 2001, Gerbig et al., 2008). Typically, surface tempera- tures are forecasted too high for calm conditions (Steeneveld et al., 2008). On the other hand, these models often show an unrealistic decoup- ling of the atmosphere from the surface, resulting in runaway surface cooling (e.g. Mahrt, 1998). The poor model representation of the SBL results in evident problems for air quality prediction (Sal- mond and McKendry, 2005). Thus improved un- derstanding and representation of the SBL is de- sirable. This paper focuses on the role of long wave radiative cooling in the heat budget of the SBL close to the surface (lowest 20 m). Our aim is to analyze and quantify long wave radiation diver- gence (LWRD) by means of observations. Finally, we propose a practical and robust parameteriza- tion for LWRD for use in atmospheric models without the need for high computational capacity. Despite its expected importance during calm conditions, LWRD in the SBL has not been stud- ied very intensively, since most attention has been paid to the turbulent transport (e.g. Cuxart et al., 2006). LWRD occurs at (sudden) changes of the temperature or concentration of absorbing gases with height. Since this occurs close to the surface, we expect LWRD to be important for the near surface SBL heat budget. A second impor- tant term is the absorption of radiation that origi- nates from the land surface. Only limited amount of LWRD observations are available. Funk (1960, 1961) and Fuggle and Oke (1976) found typically 6.6 K h -1 LWRD (cool- 1 Corresponding author address: 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 ing) between 0.5-1.5 m AGL. On the contrary, Li- eske and Stroschein (1967) found 5 Kh -1 heating layer between 1-5 m in the Arctic. Xing-Sheng et al. (1983) reported typical cooling of 2.5 Kh -1 , with the strongest cooling at the top of a shallow sur- face inversion. Unfortunately the previous studies only cover a few nights, and determined the LWRD only over a single layer. Hoch et al. (2007) showed year- round observations of LWRD over Greenland be- low 50 m, and examined the dependence of long wave cooling the on temperature gradient, sur- face humidity and wind speed. They concluded that the divergence of the outgoing long wave flux is the dominant contributor to the total LWRD. Recently, Drüe and Heinemann (2007) reported LWRD over Greenland up to 800 m. They found that LWRD is important not only during calm con- ditions, but also under moderate wind speeds. However, Estournel et al. (1986) found LWRD of secondary impact during the ECATS campaign. for stable nights with moderate winds. Sun et al. (2003) estimated LWRD between 48 and 2 m during CASES-99, and reported a small impact of LWRD during the core of the night. Moores (1982) documented observations that indicated substantial long wave heating in the daytime boundary layer. Below 150 m he found 0.05-0.52 Kh -1 heating (30% of the turbulent heating), and 0.06 K/h cooling above 150 m. Over all, we con- clude that the relative role of radiative cooling is under debate, as are the numerical values. Several technical limitations and difficulties need to be overcome (e.g. dew formation, ventila- tion, bias correction, mast influence), which ex- plains the limited amount of research. Typically measurement uncertainties are close to the re- corded signal. Amongst others, Ha and Mahrt (2003) modeled LWRD. They and Räisänen (1996) showed strong sensitivity to model resolution, and even that the forecasted sign of the radiative tendency can be wrong for coarse resolution. Also, running a full radiation scheme that accounts for all ab- sorbing gases at very grid cell and every time step is computationally too expensive for NWP and climate models. Moreover, these schemes have typically been calibrated and validated against the observed cooling in the full atmos- pheric column. However, the SBL is a very shal- low layer with strong temperature and humidity gradients, which can differ substantially from free