Status and Challenges in Nowcasting High Impact Weather Rita - - PowerPoint PPT Presentation

▶
status and challenges in nowcasting high impact weather
SMART_READER_LITE
LIVE PREVIEW

Status and Challenges in Nowcasting High Impact Weather Rita - - PowerPoint PPT Presentation

Feb 04, 2024 166 likes •526 views

Status and Challenges in Nowcasting High Impact Weather Rita Roberts National Center for Atmospheric Research With contributions from Jim Wilson WSN16 7/25/2016 Rita Roberts NCAR/RAL Nowcasting Timeline WMO Forecast Kano, Nigeria, 1956

slide-1
SLIDE 1

Status and Challenges in Nowcasting High Impact Weather

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Rita Roberts National Center for Atmospheric Research

With contributions from Jim Wilson

slide-2
SLIDE 2

Nowcasting Timeline

Herb Ligda

Extrapolation of Radar Echoes WSR-57 Radar Detection of Colliding Sea Breezes First U.S. weather satellite launched from Cape Canaveral, FL

James Purdom: Some uses of GOES imagery in mesoscale convection forecasting International Nowcasting Demonstrations WMO Forecast Demonstration Program Sydney 2000 Olympics

1953 1965 1984 1960 1976 2000

Kano, Nigeria, 1956 Microburst Aviation Accident WSN16 7/25/2016 Rita Roberts NCAR/RAL

slide-3
SLIDE 3

1984

WSN16 7/25/2016 Rita Roberts NCAR/RAL Use of satellite soundings and imagery for nowcasting and very-short-range forecasting

  • J. F. W. Purdom

A scheme for the delineation and extrapolation of thunderstorms from GOES satellite data

  • G. L. Austin and A. Bellon

Nowcasting with Doppler radar: the forecaster-computer relationship

  • J. Wilson and R. Carbone

Short-period forecasts in West Africa using Meteosat data

  • J. R. Milford and G. Dugdale

Digital radar data as an aid in nowcasting in Hong Kong

  • C. Y. Lam

The United Kingdom Meteorological Office mesoscale forecasting system

  • B. W. Golding and N. A. Machin

A one-level mesoscale model suitable for nowcasting and short-term forecasting in regions of complex terrain

  • C. F. Mass and D. P. Dempsey

FRONTIERS – Progress with a system for nowcasting rain

  • K. M. Carpenter and K. A. Browning

Interactive computer display systems in nowcasting D.A. Chisholm and A. J. Jackson Nowcasting using lightning location techniques

  • R. E. Orville

User requirements for very short range weather forecasts: Some key issues

  • A. M. Murphy and B. B. Brown

We’ve been at this for a long time!!

Are we making progress?

slide-4
SLIDE 4

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Tornadoes

Flash floods/ land slides Sporting Events Aviation

Lake Victoria fisherman

Marine Weather Road weather Transportation

What is it going to take to improve nowcasts, watches and warnings of high impact weather so that people will take action to save lives and property?

slide-5
SLIDE 5

OUTLINE

WSN16 7/25/2016 Rita Roberts NCAR/RAL

  • Status of 0-6 hour nowcasting
  • Challenges in nowcasting
  • Promising technologies and needs
slide-6
SLIDE 6

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Bamako, Mali

9 June 2008

NOWCASTING STATUS

Bamako-Senou International Airport

We know the importance of convergence boundaries for

convection Initiation, thunderstorms, wind shifts and heavy rain

slide-7
SLIDE 7

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Colorado, U.S.A. Tiwi Islands, Australia Beijing, China

Gust front

Bamako, Mali

Convergence boundaries are ubiquitous! They play a role in :

  • convection initiation
  • surface moisture gradients
  • storm growth
  • updraft and downdraft intensities
  • storm dissipation
  • upscale growth of storms into MCSs

a) 0100 UTC

b) 0900 UTC

c) 1800 UTC

Dust RGB plots of low-level convergence boundaries and storm initiation and evolution above these boundaries (Jochen Kerkmann and Daniel Rosenfeld).

NOWCASTING STATUS

slide-8
SLIDE 8

The track of the dusty gust front, using RGB dust imagery, associated with a synoptic-scale event from 3-5 August 2006, propagating northward through West Africa.

For the period from June-September 2006, the distances traveled by gust fronts ranged from 160 km – 1920 km!

  • M. McGraw-Herdeg, M.S. Thesis, 2010

Long-lived nature of gust fronts associated with MCS convection in North-Western Africa

WSN16 7/25/2016 Rita Roberts NCAR/RAL

NOWCASTING STATUS

slide-9
SLIDE 9

NOWCASTING STATUS

Radar echo extrapolation is still primary tool ( over 59 years)

Among the most used are: TITAN, SCIT, MAPLE, TREC, Optical Flow

TITAN as used in Australia

Nowcast lead time Skill

Extrapolation skill drops rapidly with time

Storm echo Time 1 Time 2 Time 3 Time 4 Nowcast for Time 5

East North

Trending size and intensity not very useful

Nowcasting Tools

WSN16 7/25/2016 Rita Roberts NCAR/RAL

slide-10
SLIDE 10

WSN16 7/25/2016 Rita Roberts NCAR/RAL

TITAN Storm tracks for 01 July 2015 During the period from 01 June – 15 Aug 2014 91% of storms that initiated above an elevation of 2600 m dissipated before reaching 1400 m

Diurnal Storm Tracks along the Colorado Rocky Mountains

Colorado Rocky Mountains High Plains

1400 m elevation

Storm extrapolation techniques over complex terrain do not take into account the impact of terrain on storm growth and decay. NCAR Autonowcaster system now accounts for impact of terrain on storm extrapolation

slide-11
SLIDE 11

Forecaster warnings

  • f flash floods

Lead times for supercell tornadoes and severe weather warnings have increased significantly

Automated wind shear warnings at airports

WSN16 7/25/2016 Rita Roberts NCAR/RAL

There have been some noteworthy nowcast successes

NOWCASTING STATUS

slide-12
SLIDE 12

Increasing Rate Continuous measurements of the Liquid Water Equivalent (LWE) of snowfall and rainfall rates used to support decisions on when to de-ice planes…

Aviation Ground De-icing Display

… what type of fluid to use, and how long to de-ice….

Radar, snow gauge and Cross-correlation tracker is used in this system

Rasmussen et al., BAMS 2001 WSN16 7/25/2016 Rita Roberts NCAR/RAL

slide-13
SLIDE 13

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Need for Integrated Displays

Met́éo France Rapid Development Thunderstorm (RDT) display

More forecast offices are setting up integrated displays for viewing nowcast products and producing watches and warnings

NOWCASTING STATUS

The message is being heard!

slide-14
SLIDE 14

Example from NCAR AutoNowcaster

Radar Reflectivity at 08Aug2008 0711PM Reflectivity one hour before flash flood Where will the flood be?

1hr nowcast of rainfall amount

Flash Flood in Denver

Expert Systems (heuristic nowcasting tools) can add improved skill over extrapolation

WSN16 7/25/2016 Rita Roberts NCAR/RAL

NOWCASTING STATUS

slide-15
SLIDE 15

Initialized 04 June 2005 00 UTC

NWP Reflectivity forecast Composite NEXRAD Radar

Real-time WRF forecast (4 km)

WSN16 7/25/2016 Rita Roberts NCAR/RAL

NOWCASTING STATUS

Limitations and Increasing Capabilities in Numerical Weather Prediction for Nowcasting

slide-16
SLIDE 16

Blending Extrapolation Forecast

Blended Forecast (12-20 UTC)

Extrap Model

Dash : 15 to 00 UTC Solid : 00 to 15 UTC

Calibration + Phase Corr

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Blending Systems

NOWCASTING STATUS

Consolidated Storm Prediction for Aviation (CoSPA)

Courtesy of James Pinto

slide-17
SLIDE 17

Nowcasting Status: Assimilation of Radar Data into NWP

07 Aug 2015 3 hr rainfall Accumulation (21-24 Z)

Radar Reflectivity, VDRAS winds at 21Z, at forecast time MRMS 3hr rain accumulation (21-24 Z)

3DVar NO Radar Assimilation 3hr forecast of rain accum (21-24 Z) 3DVar with Radar Assimilation 3hr forecast of rain accum (21-24 Z)

WSN16 7/25/2016 Rita Roberts NCAR/RAL

slide-18
SLIDE 18

WSN16 7/25/2016 Rita Roberts NCAR/RAL

1. The spin-up cycle of NWP models limits ability to use model forecasts in the 0-2hr time period. 2. Representation of background errors. 3. Insufficient observations of wind, temperature and moisture in the non-precipitating environment. 4. Insufficient techniques for extracting multi-scale information from observations and forecast background.

Limitations

NOWCASTING STATUS

Current Status of NWP with Radar Data Assimilation We asked some experts…..

Sue Ballard Jenny Sun Ming Xue Isztar Zawadzki

1) More frequent NWP update cycles and higher resolution models. 2) Reflectivity data assimilation is showing clear improvements in convective storm forecasts. Not so much with radial velocity assimilation. 3) Germany, France and UK operation centers are showing convective scale benefit with radar assimilation in 3D-Var, 4D-Var and nudging. 4) Most radar data assimilation systems have shown positive impact on hourly precipitation forecasts up to 6-8 hours (others say impact is only out to 4 hrs).

Increasing Capabilities

slide-19
SLIDE 19
  • For larger scale precipitation systems, radar data assimilation into numerical

model is improving forecasts in roughly the 3 -6 hr forecast period

  • However that improvement has not yet been realized for smaller scale

convective events.

  • For example the models may tell us that western Kansas will have an MCS,

but they are not able to forecast which community will have a flash flood and how intense it may be.

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Status in NWP Nowcasting

In summary:

slide-20
SLIDE 20

Additional barriers to improving nowcasting

Must capture boundary layer convergence lines Automated convergence line detection: Automated methods have been tried for years with only modest

  • success. The time maybe right to try again with more advanced feature

detection tools and using dual-polarization radar fields.

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Challenges in Nowcasting

slide-21
SLIDE 21
  • 50 % of nighttime convection arises from elevated

convection initiation above a stable boundary layer

  • ver the Great Plains and eastern U.S.
  • Elevated convection often organizes into MCSs,

producing heavy precipitation and flash floods

  • Specific challenges in nowcasting
  • Location and onset of initial development

Prediction of Elevated Convection Initiation

(courtesy Frederic Fabry) Rita Roberts NCAR/RAL WSN16 7/25/2016

Challenges in Nowcasting using NWP and Radar Data Assimilation

To improve scientific understanding, the Plains Elevated Convection At Night (PECAN) experiment was conducted in June-July 2015.

slide-22
SLIDE 22

Challenges in Nowcasting using NWP and Radar Data Assimilation THE PECAN EXPERIENCE

Available Model Forecasts

CSU WRF NCAR MPAS NCAR WRF ensemble NCEP GFS NCEP NAM NCEP HRRR NSSL WRF NSSL WRF ensemble OU MAP WRF OU MAP WRF ensemble mean OU MAP WRF ensemble probabilities OU MAP WRF ensemble spaghetti

Nowcaster Product

Made at 00Z MCS Forecast 03-06z 10% 30 50

WSN16 7/25/2016 Rita Roberts NCAR/RAL

slide-23
SLIDE 23

Nowcasting Timeline

Age of application discovery Browning Book and conferences Clear-air and boundary applications Sydney Olympics Beijing Olympics

Nowcasting Timeline

Echo extrapolation

WSR-57 detection of Sea Breeze

1953 1965 1960 1970’s

weather satellite

1980’s 2000 2008 1990’s

Automated Nowcasting

10 50 50 15 July 2015 06 Z Reflectivity mosaic Nowcast made 15 July 00Z

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Challenges in Nowcasting using NWP and Radar Data Assimilation

Not so good!

slide-24
SLIDE 24

Nowcasting Timeline

15 July 2015 06 Z Reflectivity mosaic Nowcast made 16Jul 00Z 16 July 2015 06 Z Reflectivity mosaic

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Challenges in Nowcasting using NWP and Radar Data Assimilation

Good!

slide-25
SLIDE 25

WSN16 7/25/2016 Rita Roberts NCAR/RAL

  • As the forecast process becomes more automated (using

NWP), the role of the forecaster in nowcasting high impact weather is still critical.

  • During PECAN, the human nowcasts were very good and

we would definitely take actions to deploy equipment were the forecasters advised. However the accuracy was not on the scale that would warrant taking life and property saving actions.

  • Many of our expert forecasters believe that hand analysis
  • f weather maps still an important step in understanding

the daily atmospheric processes in play before using NWP guidance to prepare their forecasts.

Dakar, Senegal, Aviation Forecast Desk Graduate student, PECAN field experiment

Role of the Forecaster is still critical Challenges in Nowcasting

slide-26
SLIDE 26

WSN16 7/25/2016 Rita Roberts NCAR/RAL

  • GOES-R for U.S. and S. America
  • Satellite RGB products
  • Satellite Day-Night Band and Near Constant Contrast images
  • Dual-pol radar
  • detection of hail, hydrometor classification, icing, microphysical properties
  • improved QPE for summer and winter rainfall

Promising Technologies and Activities for Nowcasting High Impact Weather

Radar-based 24-hour QPE vs gauge-measured statistics from PECAN 17 May 2015 – 16 July 2015

Method N points Correlation Bias radar/gauge NCAR HYBRID (dual-pol) 21258 0.834 0.940 NCAR Weighted-PID (dual-pol) 21311 0.826 1.108 R(Z) (single pol) 21668 0.798 1.331 R(Z, ZDR) (dual-pol) 21037 0.772 1.057

slide-27
SLIDE 27

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Using 3-D Printed Automatic Weather Stations (3D-PAWS)

  • Inexpensive
  • Built with local materials
  • Easily replaceable (if stolen or broken)
  • Compares well with reference sensors

Overcoming the lack of surface network observations

Relative humidity comparisons

Courtesy of Paul Kucera

3D-PAWS site located at Zambia Met Department Computer Aided Design (CAD) for specific instruments

Ultimate Goal:

  • obs for forecasters
  • obs for initializing regional forecast models
  • obs for verification of forecasts
slide-28
SLIDE 28

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Using Unmanned Aerial Vehicles (UAVs) to collect atmospheric information

Courtesy of Dr. Ruben Santos Universidad Nacional de Cuyo, Mendoza, Argentina

  • UAVs can be used to obtain thermodynamic profiles

in the CBL.

  • Are inexpensive compared to radiometers
  • Basic atmospheric sensors are available or can be

built Many locations in the world have limited or no soundings collected daily Ultimate Goal:

  • hourly collections of data in remote regions
  • use for initializing regional forecast models
  • collect data over complex terrain
  • for fire weather prediction
  • obs in pre-storm environments
slide-29
SLIDE 29

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Atmospheric Moisture Measurements

Micro-Pulse DIAL Lidars for Profiling Water Vapor

Courtesy of Scott Spuler, Dave Turner, Tammy Weckwerth

  • Low cost
  • Low maintenance
  • Eye safe and stable for unattended operations
  • Collect data from 300 m to 4km with 150 m resolution
  • Small instrument fits in a 2m x 2m box

Comparison to AERI profiling systems

Ultimate Goal

  • Detection of (nocturnal) atmospheric processes

above the PBL

  • Understanding and nowcasting thunderstorm

development

  • For improving and verifying NWP model moisture

fields

  • Deploy a network of instruments to obtain

continuous mapping of water vapor transport

slide-30
SLIDE 30

FAA Alaska C&V Program 30 Author: MITLL 6/1/2016

Visibility Generator

Using Web-Based Camera Imagery

Clear day composite image & edges Image Registration

Hanscom Camera 2/2/2000 through 3/9/2000 during daylight Edge Detection Image 20 40 60 80 100 120 140 100 1000 10000 Visibility (meters) Mean & Stddev (Grayscale) 2000 4000 6000 8000 10000 12000 14000 Power Mean Stddev Power

Visibility: 10 miles

Edge Detection Image Capture Normalized Edge Extraction

Courtesy of Robert Hallowell, MIT/LL

Ultimate Goal:

Retrieval of visibility conditions for aviation and road weather nowcasting. Particularly valuable in remote regions with no surface stations.

slide-31
SLIDE 31

CACTI-RELAMPAGO broad study domain

RELAMPAGO Field experiment to understand:

  • Convection initiation processes
  • Intensification and upscale growth of convection
  • Generation of hazardous weather

Argentina

  • Has one of the highest frequencies of

lightning in the world

  • Extreme flooding
  • Produces some of the largest hail in the world

Field Experiments for Scientific Understanding

Courtesy of Steve Nesbitt, U. of Illinois WSN16 7/25/2016 Rita Roberts NCAR/RAL

slide-32
SLIDE 32

Existing radar and surface stations

Additional observations for small field program

  • Mobile C-band north side
  • f lake
  • Sounding east and

west side of Lake

  • Water temperatures from

boats and buoys

  • Total cloud lightning array

Lake Victoria

Images courtesy of Burleyson and Yuter

  • Severe nocturnal thunderstorms
  • Numerous fatalities (fishermen)
  • Potentially strong wind shear
  • High lightning frequencies

Field Experiments for Scientific Understanding

slide-33
SLIDE 33

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Nowcast GOAL

Beijing Metropolitan

slide-34
SLIDE 34

WSN16 7/25/2016 Rita Roberts NCAR/RAL

Young scientists are rising to the challenge!

My 10 month old granddaughter, monitoring for wind shifts on a blade of grass...