Nowcasting Flash Floods and Heavy Precipitation --- A Satellite Perspective

1

Nowcasting Flash Floods andHeavy Precipitation ---A Satellite Perspective

---------------------------------------------

Jim Gurka

National Weather Service

Office of Meteorology

Silver Spring, MD

james.gurka@noaa.gov

301-713-1970Satmet 99-2Wednesday, 28 April 1999

2

3

4

Satellite Pictures

• GOES Water Vapor (6.7 m) day and night): detects moisture and weather systems between 700 - 300 mb

• GOES Infrared (10.5 - 12.6 m) day and night): detects cloud top temperature and surface temperature

• GOES Visible (0.55 - 0.75 m) (day only): what you can see: clouds, water, land

5

Satellite Pictures

• Polar Microwave (SSM/I and AMSU): detects precipitation, moisture, snow cover, ocean surface winds, surface wetness

6

7

Quantitative Precipitation

Forecasts (QPF)• The prediction of how much

precipitation (e.g., 1/2 inch, 1 inch, 2 inches, etc) will fall during a specified period of time (e.g., 3 hours, 6 hours, 12 hours, 24 hours, etc)

8

Quantitative Precipitation

Forecasts (QPF)• Assimilation• Analysis• Anticipation• Intuition• Experience

9

NCEP

• EMC• HPC SPC TPC MPC CPC AWC

SEC

10

Forecasting Pcpn Amounts from NWP

Models• Data Assimilation

– Sfc, Upr Air; Satellite; Radar/Profilers; ACARS•Into NWP Models

– Gives an approximate solution to wind, temperature & moisture fields in the atmosphere;

•QPF: Serves as guidance to how much precipitation will occur

11

NWP Model Output vs Observations

• 24 hours before a pcpn event:– Model output is about 17x more

useful than data;

• At time zero: Model output has virtually no value compared to observations;

• 9 hours before an event: Obs and model output have about equal value;

12

QPF: 2 Problems

•Interpreting the Numerical Guidance;

•Knowing a big rainfall event from a small one (in terms of scale and intensity)

13

Why Models Have Forecast Problems

• Lack of data over oceans/Mexico– Sparse upper air obs: lose details

• Initialization and QC…smooth data fields

• Smoothed terrain• Atmospheric process are non-

linear• Model Physics are

approximations

14

HPC Tips for Doing a QPF

• Look at precipitation history;• System evolution;• Basic ingredients & trends• Gridded data…be careful

– compare models for continuity and known biases

• Consider local effects;• Watch out for High Bias;• Verification• Warm Season: lower

expectations• Call NCEP for guidance/empathy

301-763-8343

15

QPF Improved with Resolution and Data

Assimilation• Meso-Eta, Cool Season:

– Performance better in cool season– Resolves intense thermal gradients

• Meso-Eta, Warm Season– A problem with MCS

•in the prediction of outlow boundaries

• Meso-Eta over High Terrain– low bias due to model resolution (32

km)– Betts Scheme needs adjustment for

shallow conv.

• RUC: 3 hour pcpn forecasts more accurate than from Eta

16

Model Resolution

•AVN: 110 km / 28 layers•NGM: 80 km / 16 layers•RUC: 60 km / 26 layers•Early Eta: 48 km / 38 layers•Meso-Eta: 32 km / 45 layers •Meso-Eta: 17 layers below

850 mb vs.•NGM: 3 layers

17

Data Assimilation: Goals

•Reduce “spin up” problem• Improve model QPF• Improve soil moisture

initialization• Include mesoscale

information found in precipitation data

18

Pattern Recognition

•Heavy rainfall events can often be identified with particular patterns

•Patterns are identified by:–conventional data–model output–satellite data– radar

19

Pattern Recognition

•Thickness patterns•70% saturation thickness•Preferred thickness

20

Rules of Thumb for Hvy Rain

•Max: Strongest inflow + Boundary

•Max: Just NE of Theta-E Ridge

•Along outflow S of Wmfnt (summer)

• Inverted isobars along a front

•MCS track alng or Rt of K lines

•Convection Srn edge of Westerlies

•MCSs often form near upr ridge

21

Rules of Thumb for Hvy Rain

•Beware slow moving synoptic systems (SHARS) …often have warm tops

•MCSs often form near upr ridge

•Strong hgt falls / fast movg systems usually preclude very heavy rafl

•Models usually predict axis of heaviest rain too far north (outflows)

22

Rules of Thumb for Hvy Rain

•K index: good measure of deep moisture–beware Ks in upper 30s

•Tropical Systems Max rainfall–usually in core at night–usually along periphery in daytime

23

To Release CAPE:

•Syn scale forcing does not act quickly enough (but moistens air & wkns cap)

•Need meso source of lift to reach LFC– low level boundaries/ fronts–find boundaries in T, DP, Theta-E & wind

– low level convergence

24

Importance of CIN

•Steep lapse rate above inversion–classic loaded gun sounding

•Cap stores energy leading to higher potential buoyant energy later in the day or evening

•Slightly stronger CIN upstream can sometimes lead to backbuilding

25

Eta Characteristics

•Most srn bias for cyclone track–esp east of Appalachians

•Too slow and too far west with strng surface cyclones coming up East coast

•Sometimes overdevelops LLJ•Too much Pcpn Gulf and Atl

coasts–due to diff in convective parameterization

•Vort centers too strong (with convct)

26

Eta Characteristics

•Significantly underforecasts monsoonal convct over SW states

•Best at fcstg frntl psn (using LI grdnt) in damming events ovr E. U.S.

•Often slightly overforecasts strength of Arctic anticyclones

•Not enough cnvctn ovr plains (summ)

•Too wet in Florida

27

AVN Characteristics

•Cold bias (mainly E. U.S.)•Genly 1st short range model

to fcst cyclogenesis–can be too slow deepening

•Best with filling cyclones•Best psn fcst cyclones and

anticy (days 2 & 3)– too far N & W sfc cycl formg E of Rockies

–once upr trof…lee Rockies…OK

28

AVN Characteristics

•Often weakens Srn stream system or phases the 2 separate systems;

•Too slow…Arctic airmass in the plains

•Overdoes upslp pcpn Cntrl and Srn Rockies and strength of upslope flow

•Oct-April Avn QPF quite good–competitive with Eta and btr than NGM

29

AVN Characteristics

•With strong wrapped up systems, forecasts hvy pcpn too far N & W

•Tracks tropical systems too far W

•Overdeepens weak tropical waves

•Overpredicts the # and strength of anticyclones ovr Grt Basin (cool ssn)

•Significantly underpredicts monsoonal precipitation

30

HPC QPF Summary

•Start with an accurate assessment of the atmosphere

•Pattern Recognition

•Know the model biases

31

32

Questions to ask when preparing Nowcasts and

QPF“this involves”

•Satellite interpretations / signatures

•Satellite conceptual models

•satellite products

33

Global to Synoptic Scale

•Wetness of ground?•Pattern recognition?

34

Soil (surface) Wetness Index (85 GHz - 19 GHz) (H)for January 1995

35

Soil (surface) Wetness Index (85 GHz - 19 GHz) (H)5 day moving averages ending August 7, 1998

36

PATTERN RECOGNITION: 6.7 micron water vapor imageryfor 11-26-97 0000 UTC; 300 mb winds (mps) are superimposed

37

PATTERN RECOGNITION: 6.7 micron water vaporimagery for 6-11-98 1200 UTC; 300 mb winds (mps)are superimposed

38PATTERN RECOGNITION: 6.7 Micron for 2-3-98 1200 UTC;300 mb winds (mps) are superimposed

39

Synoptic to Mesoscale

• Available moisture and stability?• Presence of lifting or lid?• presence of boundaries?

(synoptic of mesoscale)

40

Use of Water Vapor(6.7 m) Imagery

• middle - upper level flow fields and circulations

• lifting mechanisms- jet streaks

- cyclonic circulations/lobes - trough axes - mid-level cold air advection

• excellent data for pattern recognition

41

6.7 m Water Vapor Plumes

• Makes environment more efficient?

• Enhances precipitation through cloud seeding?

• associated with favorable synoptic patterns for low-level moisture and instability

42

WATER VAPOR PLUME: 6.7 micron water vaporimagery for 10-16-98 2345 UTC; 300 mb winds (mps)are superimposed

43

WATER VAPOR PLUME: 6.7 micron water vaporimagery for 10-17-98 1215 UTC

446.7 micron water vapor imagery for 6-22-97, 1745 UTC

45

46

Precipitation Efficiency Factors

• Precipitable Water (PW) values ---- higher than 1.0 inch, enhances Precipitation Efficiency

• Mean environmental Relative Humidity (RH) ---- higher than 65 % results in less dry air entrainment into cloud masses

47

Precipitation Efficiency Factors

• Depth of cloud with temperatures warmer than 0 degress C enhances the collision-coalescence process by increasing residence time of droplets in clouds --- this increases rainfall intensity and improves precipitation efficiency

48

Precipitation Efficiency Factors

• Vertical wind shear ---- produces entrainment and reduces Precipitation Efficiency, especially if environmental air is dry

• Cloud-scale vertical motion function of “CAPE” (Convective Available Potential Energy) related to condensate production and residence time of droplets

49

Additional Precipitation

Efficiency Factors• Storm-relative mean inflow and

moisture transport into storm• Duration of precipitation

50

Use of Precipitable Water (PW) and

Relative Humidity (RH) Data

• magnitude• transport (plumes)• trends• relation to equivalent potential

temperature

51

PW Plume: Composited PW (mm) (SSM/I + GOES + ETA model)for 10-17-98 1200 UTC; 850 mb winds are superimposed

52PW Plume: Composited Precipitable Water Product (mm) for6-22-97 1800 UTC; 850 mb winds superimposed

53

Precipitable Water Available

• SSM/I (polar microwave)- water only

• GOES 8/9/10- clear (cloud free) areas

• National Center for Environmental Prediction (NCEP) “ETA” and “AVN” models

• Rawinsondes• Composites (SSM/I + GOES 8/9/10

+ ETA/AVN)

54

Defense MeteorologicalSatellite

Program/SpecialSensor Microwave

Imager(SSM/I) Channel most

sensitive to Total Precipitable Water

22.235 GHz

55

GOES 8 Sounding Channels Sensitive to Precipitable Water (m) (over land and water, cannot calculate when clouds are present)

14.37; 14.06; 13.96; 13.37; 12.66; 12:02; 11.03; 7.43; 7.02; 6.51; 4.57; 4.52; 4.13

Cloud Detection

3.74 and 0.94

56GOES Precipitable Water Products (mm) for October 22 - 23, 1995

57

Precipitable Water X Relative Humidity

• Adjusts satellite-derived Quantitative Precipitation Estimates (QPE)

• Used in Quantitative Precipitation Forecasts (QPF) Techniques

58

Precipitable Water (mm) X Relative Humidity for10-17-98 1200 UTC from the ETA Model; 850 mb winds (mps)are superimposed

59

Equivalent Potential Temperature Ridge

Axis• Represents a potential energy

axis for convective development• Conservative (similar to vorticity)• “Maximum areas” are often

associated with afternoon to evening convection

60

Equivalent Potential Temperature Ridge

Axis• “Gradient areas” (especially

north of the ridge axis) are often associated with “overrunning” nocturnal convection

61850 mb equivalent potential temperature (degrees K) for 10-17-98 1200 UTC

62

6.7 micron water vapor imagery for 6-20-96 0515 UTC

65

GOES 8/9 Precipitable Water Product (mm) for 6-20-96 0546 UTC

66850 mb winds (kts) for 6-20-96 0000 UTC

67850 mb equivalent potential temperature (degrees K)for 6-20-96 0000 UTC

68

69

Short Waves(lifting mechanisms)

• Dark areas detected in the 6.7 m water vapor imagery:

- mid-level cold air advection - jet streaks

- positive vorticity advection

- trough axes- height fall (rise)

centers

706.7 micron water vapor imagery for 6-26-95 1915 UTC