Land Surface Processes and Their Impact in Weather Forecasting

Transcription

1 Land Surface Processes and Their Impact in Weather Forecasting Andrea Hahmann NCAR/RAL with thanks to P. Dirmeyer (COLA) and R. Koster (NASA/GSFC) Forecasters Conference Summer 2005 Andrea Hahmann ATEC Forecasters Conference 2005 Slide 1

2 Outline Land surface processes Budgets, budgets, and more budgets How are these processes represented in numerical models? What land surface processes are important for weather (and climate)? What is a land data assimilation system? Why do we need it? Andrea Hahmann ATEC Forecasters Conference 2005 Slide 2

3 Outline Land surface processes Budgets, budgets, and more budgets How are these processes represented in numerical models? What land surface processes are important for weather forecasting? What is a land data assimilation system? Why do we need it? Andrea Hahmann ATEC Forecasters Conference 2005 Slide 3

4 Land surface processes Land surface processes are those processes that occur at the interface between the land surface and the atmosphere. Land surface processes are important to numerical weather prediction because they provide lower boundary conditions to atmospheric physical and dynamical processes. The land surface is a source/sink of energy, water (and other constituents such as carbon dioxide, nitrogen), and momentum to the atmosphere. Andrea Hahmann ATEC Forecasters Conference 2005 Slide 4

5 In the ATEC RTFDDA These processes are simulated by a land surface model called Noah (NCAR, OSU, AirForce, office of Hydrology) LSM We will now try to understand how these processes are represented in this model Andrea Hahmann ATEC Forecasters Conference 2005 Slide 5

6 Noah LSM in NCEP ETA, MM5 and WRF (Pan and M a h r t, 1987; Chen et al., 1996; Chen and Dudhia, 2001; Ek et al., 2003) Canopy Water Transpiration Evaporation Turbulent Heat Flux to/from Snowpack/Soil/Plant Canopy Precipitation Condensation Runoff on vegetation on bare soil Direct Soil Evaporation Deposition/ Sublimation to/from snowpack Evaporation from Open Water Snowmelt = 10 cm Soil Moisture Flux Interflow Soil Heat Flux = 30 cm Internal Soil Moisture Flux = 60 cm Internal Soil Heat Flux = 100 cm Andrea Hahmann Gravitational Flow ATEC Forecasters Conference 2005 Slide 6

7 The surface energy balance The change in temperature of the surface is largely determined by the balance between incoming and outgoing surface energy Andrea Hahmann ATEC Forecasters Conference 2005 Slide 7

8 The surface energy balance equation Terms on LHS come from the atmospheric model. Strongly dependent on cloudiness, water vapor, etc. Energy balance for a single land surface slab, without snow S w S w L w L w H LE T S w + L w = S w + L w + H + LE + C p T + miscellaneous where S w = Incoming shortwave radiation L w = Downward longwave radiation S w = Reflected shortwave radiation L w = Upward longwave radiation H = Sensible heat flux L = latent heat of vaporization E = Evaporation rate C p = Heat capacity of surface slab T = Change in slab s temperature, over the time step miscellaneous = energy associated with soil water freezing, plant chemical energy, heat content of precipitation, etc. Terms on RHS are determined by the land surface model. Andrea Hahmann ATEC Forecasters Conference 2005 Slide 8

9 Other energy balances are also considered. For example: Energy balance of a vegetation canopy Energy balance in a surface layer S w S w L w L w H LE Note: same symbols are used, but values will be different. S w S w L w L w H LE T c Internal energy T 1 T 2 G 12 S w S w L w L w H H T 3 G 12 = heat flux between soil layers 1 and 2 Energy balance in a subsurface layer Energy balance in snowpack S w S w L w L w H LE T 1 G 12 Internal energy T snow G S1 l m M Internal energy T 2 G 23 T 1 T 3 m = latent heat of melting s = latent heat of sublimation M = snowmelt rate G S1 = heat flux between bottom of pack and soil layer 1 Andrea Hahmann ATEC Forecasters Conference 2005 Slide 9

10 The surface energy balance (cont) In the example below, a total of five energy balances are computed: one for the canopy, one for the snowpack, and one for each of three soil layers. Note that some models may include additional soil layers or may divide the snowpack itself into layers, each with its own energy balance. Andrea Hahmann ATEC Forecasters Conference 2005 Slide 10

11 Radiation components Reflected solar radiation: Simplest description: consider only one band (the whole spectrum) and do not differentiate between diffuse and direct components: S w = S w surface albedo (reflectance) Typical albedos for various surfaces Sand Grassland Green crops Forests Dense forests Fresh snow Old snow Urban Upward longwave flux: Stefan-Boltzman Law: L w = T s 4 Surface emissivity ~1 in most models Stefan- Boltzman constant = 5.67 x 10-8 W/(m2.K4) Andrea Hahmann ATEC Forecasters Conference 2005 Slide 11

12 The diurnal cycle of surface fluxes: atmospheric forcing Time series at a point incident shortwave (solar) radiation - strong diurnal cycle SW rad. DPG D3, 13:00 LST downward longwave (from atmosphere) radiation - uniform through the day Andrea Hahmann ATEC Forecasters Conference 2005 Slide 12

13 Surface fluxes Sensible heat flux: air density H = c p C H V (T s -T r ) specific heat of air exchange coef. for heat wind speed at ref. level surface temperature exchange coefficients depend on atmospheric stability, roughness of the surface. air temperature at ref. level Andrea Hahmann ATEC Forecasters Conference 2005 Slide 13

14 Surface fluxes (cont.) latent heat of vaporization, sublimation, or melting evaporation rate Latent heat flux: Simplest form: Coeff. depending on soil moisture, vegetation exchange coef. for water vapor saturated LE = L C DE V [q * (T s )-q r ] surface specific humidity _beta soil moisture ratio In reality there are four evaporation components: transpiration bare soil evaporation interception loss snow evaporation Andrea Hahmann ATEC Forecasters Conference 2005 Slide 14

15 Soil moisture and its impact on the energy budget DPG, D3, volumetric soil moisture in layer 1 H LE LE: Latent Heat Flux LE H H: Sensible Heat Flux Andrea Hahmann ATEC Forecasters Conference 2005 Slide 15

16 Soil moisture and its impact on the energy budget (and temperature) DPG, D3, volumetric soil moisture in layer 1 T2 TG TG T2 T2: 2-m air Temp. TG: Ground Temp. Andrea Hahmann ATEC Forecasters Conference 2005 Slide 16

17 In summary The surface energy balance under varied soil wetness conditions Andrea Hahmann ATEC Forecasters Conference 2005 Slide 17

18 Energy balance of the snowpack Albedo is high when the snow is fresh, but it decreases as the snow ages. Energy balance in snowpack Snowmelt occurs only when snow temperature reaches o K. Internal energy a function of snow amount, snow temperature, and liquid water retention Solid fraction 1 0 S w S w L w L w H LE Internal energy Temperature T snow T 1 G S1 LM Thermal conductivity within snow pack varies with snow age. It increases with snow density (compaction over time) and with liquid water retention. Andrea Hahmann ATEC Forecasters Conference 2005 Slide 18

19 The surface water balance W in Control volume water storage W out The soil water content in the layer closest to the surface is mainly determined by the balance between precipitation and evaporation Water within the soil is distributed by drainage and diffusion Andrea Hahmann ATEC Forecasters Conference 2005 Slide 19

20 Water Balance for a single land surface slab, without snow Terms on LHS come from the atmospheric model. Strongly dependent on cloudiness, water vapor, etc. P P E R w = E + R + C w w/ t + miscellaneous Terms on RHS come are determined by the land surface model. where: P = Precipitation E = Evaporation R = Runoff (effectively consisting of surface runoff and baseflow) C w = Water holding capacity of surface slab w = Change in the degree of saturation of the surface slab t = time step length miscellaneous = conversion to plant sugars, human consumption, etc. Andrea Hahmann ATEC Forecasters Conference 2005 Slide 20

21 Two other water budgets Water balance associated with canopy interception reservoir P D c W c E int P = E int + D c + W c t E int = interception loss D c = drainage through canopy ( throughfall ) W c = change in canopy interception storage Water balance in a snowpack P (snow) E snow P = E snow + M + W snow t M W snow E snow = sublimation rate M = snowmelt W snow = change in snow amount ( infinite capacity possible) Andrea Hahmann ATEC Forecasters Conference 2005 Slide 21

22 Energy balance vs. water balance Energy balance: Implicit solution usually necessary Results in updated temperature prognostics Water balance: Implicit solution usually not necessary Results in updated water storage prognostics How are the energy and water budgets connected? 1. Evaporation appears in both 2. Albedo varies with soil moisture content 3. Thermal conductivity varies with soil moisture content 4. Thermal emissivity varies with soil moisture content Andrea Hahmann ATEC Forecasters Conference 2005 Slide 22

23 Important component of a land surface model Distribution of land cover and soil types (map) Look-up table for vegetation, soil parameters JViz demo Andrea Hahmann ATEC Forecasters Conference 2005 Slide 23

24 Distribution of land cover types Similar map for soil types Andrea Hahmann ATEC Forecasters Conference 2005 Slide 24

25 Land cover parameters (Noah LSM as coupled to MM5 or WRF) Land cover Albedo (%) Roughness length (m). Land cover type Urban and build-in land Dryland cropland and pasture Irrigated cropland and pasture. ATEC only Playa Lava White sand similar look-up table for soil types Andrea Hahmann ATEC Forecasters Conference 2005 Slide 25

26 Outline Land surface balances Budgets, budgets, and more budgets How are these processes represented in numerical models? What land surface processes are important for weather and climate? What is a land data assimilation system? Why do we need it? Andrea Hahmann ATEC Forecasters Conference 2005 Slide 26

27 Land-atmospheric feedback on precipitation (precip -> soil moisture) Precipitation wets the surface... causing soil moisture to increase... which affects the overlying atmosphere (the boundary layer structure, humidity, etc.)... which causes evaporation to increase during subsequent days... thereby (maybe) inducing additional precipitation Andrea Hahmann ATEC Forecasters Conference 2005 Slide 27

28 Land-atmospheric feedback on precipitation (precip, soil moisture) Precipitation wets the surface... causing soil moisture to increase (more clouds, less sunshine)... which affects the overlying atmosphere; more stable atmosphere which causes temperature to decrease... thereby (maybe) inhibiting additional precipitation Andrea Hahmann ATEC Forecasters Conference 2005 Slide 28

29 Land-atmospheric feedback on precipitation (gradients ) Rains on the left but not on the right causing wetter soil on the left than on the right which causes a horizontal temperature gradient... which affects the horizontal pressure gradient thereby inducing mesoscale circulations (which produce rainfall) cooler warmer wet dry Tleft < Tright Andrea Hahmann ATEC Forecasters Conference 2005 Slide 29

30 Outline Land surface balances Budgets, budgets, and more budgets How are these processes represented in numerical models? What land surface processes are important for weather and climate? What is a land data assimilation system? Why do we need it? Andrea Hahmann ATEC Forecasters Conference 2005 Slide 30

31 HRLDAS example: soil moisture Aberdeen Test Center (ATC), D1 Time series of domain-averaged volumetric soil Snap shots of daily (12Z) volumetric moisture soil (top moisture, 3 layers) surface and precipitation (12hrs) layer, July 20-31, 2005 high spatial variability high temporal variability 0-10 cm cm cm Andrea Hahmann ATEC Forecasters Conference 2005 Slide 31

32 Why use land data assimilation? Now that we understand the importance of land surface processes (and their spatial variability) in weather forecasting Mesoscale models need to capture atmospheric boundary layer structures and motions resulting from these surface heterogeneities But, there are no routine high-resolution soil observations at continental scale available for mesoscale coupled system initialization Andrea Hahmann ATEC Forecasters Conference 2005 Slide 32

33 Land Data Assimilation System: Provides a continuous record of the state of the land surface Andrea Hahmann ATEC Forecasters Conference 2005 Slide 33

34 Domain-averaged soil moisture: DPG Domain 2, 01/03-04/ Total volumetric soil water (liquid + frozen) Liquid soil water Andrea Hahmann ATEC Forecasters Conference 2005 Slide 34

35 Comparison of initial soil moisture field, DPG, valid April 2, :00 UTC DPG D1 DPG D3 Initial conditions provided by Andrea Hahmann ATEC Forecasters Conference 2005 Slide 35

36 Land data assimilation system (cont) Biases in energy and water stores can develop in coupled modeling systems (just reuse land surface fields from one model cycle to the next) due to forcing errors, and errors in model physics and parameters, and continue to grow in such self contained systems Uncoupled Land Data Assimilation Systems (LDAS) driven by observations and constrained by data assimilation have potential to more accurately depict land surface conditions Andrea Hahmann ATEC Forecasters Conference 2005 Slide 36

37 Coupled land data assimilation system Andrea Hahmann ATEC Forecasters Conference 2005 Slide 37

38 Comparison of water equivalent snow cover depth SNOW Analysis (NOHRSC) MM5 model :00 UTC cycle HRLDAS :00 UTC cycle Andrea Hahmann ATEC Forecasters Conference 2005 Slide 38

39 In summary Correct representation of land surface processes is important to an accurate representation of physical and dynamical processes near the earth surface (and above); In the ATEC RTFDDA system, MM5 contains such processes in a model named Noah LSM; These process are spatially heterogeneous and this variability is important to a correct representation of the processes in the boundary layer; Since no soil moisture observations are available, a land data assimilation system (HRLDAS) is used to initialize surface-related quantities (DPG GMOD only for now) HRLDAS web page: Andrea Hahmann ATEC Forecasters Conference 2005 Slide 39