Convective Parameterization in NWP Models Jack Kain And Mike Baldwin What is convective parameterization? A technique used in NWP to predict the collective effects of (many) convective clouds that may exist within a single grid element…As a function of larger-scale processes and/or conditions. Why do NWP models need to worry about it? Direct Concern: To Predict convective precipitation Feedback to larger Scales: Deep convection “overturns” the atmosphere, strongly affecting mesoscale dynamics - Changes vertical stability - generates and redistributes heat - removes and redistributes moisture - makes clouds, strongly affecting surface heating and atmospheric radiation A convective parameterization must decide 3 things: • Activation? Trigger function • Intensity? Closure Assumptions • Vertical Distribution? Cloud model or specified profile Trigger Functions Cloud C Moist. Sub-cloud Cloud-layer CAPE Depth I Conv. Mass conv. Moisture ∂(CAPE)/∂t N BMJ (Eta) Grell (RUC, AVN) KF (Research) Bougeault (Meteo FR) Tiedtke (ECMWF) Bechtold (Research) Emanuel (Research) Closure Assumptions (Intensity) CAPE Cloud-layer Moisture ∂(CAPE)/∂t Subcloud moisture Converg. Quasi-equil. BMJ (Eta) Grell (RUC, AVN) KF (Research) Bougeault (Meteo FR) Tiedtke (ECMWF) Bechtold (Research) Emanuel (Research) Vertical Distribution of Heat, Moisture Entraining/Detraining Convective Buoyancy Plume Adjustment Sorting Cloud Model BMJ (Eta) Grell (RUC, AVN) KF (Research) Bougeault (Meteo FR) Tiedtke (ECMWF) Bechtold (Research) Emanuel (Research) How is the parameterized information fed back to the model? Consider the Temperature-Tendency Equation in a model: d Prad Pconv Pcond / evap Phmix Pvmix Psfc dt Where the convective term is simply adj init Pconv t conv c Consider two very different approaches: 1) BMJ Scheme (convective adjustment) 2) KF scheme (mass flux scheme) Procedure followed by BMJ scheme… 1) Find the most unstable air in lowest ~ 200 mb 2) Draw a moist adiabat for this air 3) Compute a first-guess temperature- adjustment profile (Tref) 4) Compute a first-guess dewpoint- adjustment profile (qref) The Next Step is an Enthalpy Adjustment First Law of Thermodynamics: dH C p dT Lv dqv With Parameterized Convection, each grid-point column is treated in isolation. Total column latent heating must be directly proportional to total column drying, or dH = 0. C p Tref T dp Lv (qvref qv )dp Pt Pt Pb Pb Enthalpy is not conserved for first-guess profiles for this sounding! Must shift Tref and qvref to the left… Imposing Enthalpy Adjustment: Adjusted Enthalpy Profiles: Suppose the cloud layer was drier…reduce RH by 15%: Enthalpy is conserved, but the net temperature change is negative, and the net moisture change is positive: Negative Precipitation! If we systematically change cloud-layer RH in this sounding, it can be shown that precipitation rate generated by the scheme is very sensitive to deep-layer moisture: If the environment is too dry or CAPE layer is less than ~ 200 mb deep, the scheme attempts to initiate shallow (non-precipitating) convection 1) Set cloud-top height as the level within 200 mb of LCL where RH falls off most rapidly with height. 2) Find LCL of cloud-top air; line connecting LCLs of subcloud and cloud-top air is a “mixing line”. 3) Assume Tref has same slope as mixing line; first-guess Tref is anchored on ambient temperature curve. With Shallow Convection, there is no net temperature or moisture change: Pt Pt Pb C p (Tref T )dp 0 and Pb Lv (qvref qv )dp 0 Consider the impact of parameterized BMJ shallow convection in a “normal” diurnal cycle… Model Initial Condition Raob BMX 12 Z 11 May 2000 Convective Adjustment Profiles… Initial time 1 h forecast Convective Adjustment Profiles… 6 h forecast: BMJ convection inactive because “convective entropy change” would be negative. 3 h forecast Sounding characteristics that lead to negative entropy change are not easily identified. Other constraints that cause BMJ shallow convection to “abort”: - Net entropy change in cloud layer would be negative - Tref is super-adiabatic - Tref is isothermal - qref gives a negative q at some level - qref gives an increase in q with height - qref gives super-saturated q at some level Back to the convective adjustment profiles… 9 h forecast – 2100 UTC Compare with raob at 00 Z: 12 h forecast Model forecast Raob BMX 00Z 12 May 2000 Consider a transition from shallow to deep convection… FWD 00Z 20 April 2001 Model Initial Condition Raob Convective Adjustment Profiles… 1h Forecast Compare with raob at 12 Z: 12 h forecast Model forecast Raob FWD 12Z 20 April 2001 More Convective Adjustment Profiles… 16 h forecast 17 h forecast Continuing to work on the sounding… 18 h forecast 21 h forecast Compare with raob at 00 Z: 24 h forecast Raob Model Forecast FWD 12Z 20 April 2001 BMJ Deep convection activated only briefly at FWD, but 100 miles to the north (ADM), BMJ deep convection was more persistent and strongly modified soundings: EtaKF Model Forecast Model Forecast ADM 22 Z 20 April 2001 OK, consider the KF scheme, a “Mass-flux” parameterization Basic procedures… 1) Starting at the surface, mix ~ 50 mb deep layers, lift to LCL 2) Give parcel a boost based on low-level convergence. Can it reach the LFC? 3) If parcel makes it to LFC, allow it to rise and overshoot equilibrium level. 4) Form downdraft from air within ~ 200 mb of cloud base 5) Overturn mass in updraft, downdraft, and surrounding environment until stabilization is achieved. If cloud depth 3 km, parameterize shallow convection Updraft Source Layer KF adjustment profiles Focus on deep convection…what is the Updraft Mass Flux (UMF*)? The mass of air that goes through cloud base divided by the initial mass in the ~ 50 mb updraft source layer: UMF* = Mu/Musl How is UMF determined? What is UMF* sensitive to? e of downdraft air Lapse rates in cloud layer Increasing humidity in the 900 – 550 mb layer increases downdraft e. This makes stabilization of the boundary layer less efficient and UMF* increases. Summary Parameterized shallow convection can distort sounding structures, significantly affecting CIN and CAPE; more problematic with BMJ than with KF BMJ deep convection very sensitive to cloud-layer RH KF mass flux particularly sensitive to lapse rates in lower half of cloud layer.