Radiation Fog Forecasting Using a 1-D Model Lionel Peyraud MétéoSwiss – Geneva Switzerland

Possible Solutions 3-D mesoscale numerical forecasting (i.e. MM5) Statistical regression forecasting (i.e. MOS) Link between observations/model forecast and visibility based on archived cases in database. Most statistically significant predictors retained Expert systems, various empirical rules and conceptual models. And also… High resolution 1-D forecasting of the boundary layer -MM5: Current finest resolution with model ~ + large computer time necessary with added finer vertical rez. - MOS: Model Output Statistics currently has obstruction to visibility forecasts including fog.

Objectives To establish a baseline performance for the COBEL 1-D model for radiation fog forecasting Get model simulations in close agreement with actual weather observations. Tune the model so that the simulations portray realistic fluctuations of weather parameters with respect to the incorporated model parameterizations.

Objectives Interface COBEL with various 3-D high resolution numerical models and determine optimal coupled configuration. Establish a list of weather parameters that affect fog the most and attribute to them a range that span the observed forecasting outcomes obtained through the One-Dimensional Ensemble Technique (ODEP). Incorporation of an assimilation technique utilizing relevant instrumentation at airport sites.

Objectives Run the coupled COBEL model in a semi-operational mode. Synthesize results and establish a useful forecasting scheme with revisions as necessary. Run coupled COBEL model as a real-time forecasting tool.

A Little History on COBEL… Developed in France for the study of physical processes in nocturnal BL (Univ. Paul Sabatier, 1988) Used as a fog forecasting tool (1993) Paul Sabatier + Météo France (Nord pas de Calais) Adapted to day conditions, more external forcings, data assimilation, wake vortex studies (UQAM, 1993-1997) Stratus dissipation forecasting at SFO (UQAM, 1997 - present) Radiation fog forecasting for Ohio River Basin Texas A&M + ILN NWSFO (1999 - present)

A Little History on COBEL… Low Ceiling & reduced visibility study in NYC area (NCAR, 2001 - present) Various cobel projects in France, Belgium (ongoing)

Methodology Based on: sounding data input into 1-D model integrated through time. Data 12Z vertical profiles from Payerne, CH (Temp, q, u & v wind components) Soil data (temp & moisture) Hard-coded data obtained from various sources External forcings (advections,clouds over grid, etc…) from coupled 3-D model

Methodology Local modification of variables for forecast site Soundings, soil temps, soil moisture Assimilation technique using ceilometers and sodars

COBEL Model Time Steps Dry conditions: 60 seconds Cloudy conditions: 30 seconds IR radiative calculations: 15 minutes SW radiative calculations: 1 minute Output frequency: 1 minute

Hard coded input variables into COBEL Soil moisture (2 layers=>0-5cm, 5cm-1m) Incoming solar flux at top of radiative grid (5.3km) Soil temperatures (3 levels=>0cm, -10cm, -1m) Roughness lengths (momentum & heat) Coriolis parameter Soil specific heat Surface emissivity Surface albedo Soil type Surface atmospheric pressure Vegetation parameters (currently not validated)

1-Dimensional Modeling 1-D (Column) Boundary Layer Model A Detailed Look at the Physical Processes High Vertical Resolution Radiative Transfer Turbulent Mixing Soil-Atmosphere Interactions - momentum - heat - humidity

Simulated Physical Processes LW radiative transfer (emission, absorption) SW radiative transfer (scattering, transmission, reflection, absorption) Turbulent mixing w/ variable stratification (very stable, stable, neutral, unstable profiles) Surface/atmosphere exchanges (heat, moisture, momentum) Soil moisture vertical transport (diffusion, conductivity) Diabatic effects (condensation, evaporation) Precipitation physics (autoconversion, collection, evaporation)

Physical Processes (continued) Gravitational settling of cloud droplets TKE production (by wind shear, buoyancy, transport, dissipation) Horizontal pressure force (external forcing) Horizontal advection (external forcing) (temperature, humidity, momentum) Vertical advection (by mesoscale vertical motion) Pressure tendency (external forcing)

COBEL Vertical Resolution 2 staggered grids Log-linear increase in resolution (0-1.4km) Finest: 0.5 meters Coarsest: 30 meters Above 1.4km=> 5 levels for radiative calc. Only Secondary grid=>radiative fluxes, turbulent fluxes, TKE (31 levels) Primary grid =>temp, wind, humidity, cloud water, TKE flux (30 levels) “Extension of primary grid” (5 levels=> soil temp only)

Boundary Conditions Fluxes Top: turbulent fluxes of θ, q, ql = 0 (above 1.4km), IR flux from clouds above rad. grid Bottom: flux of TKE=0 (z0), sfc albedo & emissivity Temperature Soil temp assumed constant at –1meters Wind Top: no boundary condition used Bottom: u=v=0, TKE flux=0

Initialization Above ground Observed vertical profiles of T, q, u&v wind from z=0-5.3 km Tweaked soil temps, soil moisture and lower sounding winds Altitude of stratus/fog base and top obtained via ceilometers and sodars Below ground Soil temperatures at 0cm, 10cm, 1m Soil moisture between 0-5cm, 5cm-1m

Sensitive Parameters with Ranges Parameter Range 21 June 25 August Soil Temperature 0cm=> 24-29C 0cm=>25-26C -10cm=> 23-25C -10cm=>18-21C -100cm=>14-17C -100cm=>17-20C Soil Moisture 0.400-0.476 m3/m3 0.350-0.400 m3/m3 Initial RH 61%-69% 60%-75% Low-level winds 0-5 m/s 0-4 m/s Dew deposition 70-100% 70-100% Sfc. emissivity 0.90-1.00 0.90-1.00

Less sensitive weather parameters Soil heat capacity Surface atmospheric pressure Surface albedo Soil type Roughness length

1-D Ensemble Prediction Current Method Proposed ODEP

Conclusions The COBEL 1-D model has been rather successful at reproducing fog events in the past and we are rather optimistic about its present/future applications in various locations. Some improvements are needed. Gravitational settling scheme Include vegetation parameters Soil model coupling process Coupling the model to various high resolution 3-D models and the usage of an assimilation technique should provide interesting results.

Thank You http://people.sca.uqam.ca/~tardif/COBEL/cobel_enter.htm http://www.unibas.ch/geo/mcr/3d/meteo/nmm4/nmm4.htm