NOAA Seasonal Drought Outlooks: Status and Progress Douglas Le Comte SRH/ERH Climate Workshop Peachtree City, GA June 2007.

NOAA Seasonal Drought Outlooks: NOAA Seasonal Drought Outlooks:

Status and ProgressStatus and Progress

NOAA Seasonal Drought Outlooks: NOAA Seasonal Drought Outlooks:

Status and ProgressStatus and Progress

Douglas Le ComteDouglas Le Comte

SRH/ERH Climate WorkshopPeachtree City, GA

June 2007

OutlineOutline

• Current Techniques for Drought Current Techniques for Drought ForecastingForecasting

• A closer look at the tools used to forecast A closer look at the tools used to forecast the Southeast Droughtthe Southeast Drought

• Ideas for the future of drought forecastingIdeas for the future of drought forecasting

New Forecast ToolNew Forecast Tool

• Cambodian royal cows predict drought, Cambodian royal cows predict drought, poor harvest this year (poor harvest this year (2007/5/62007/5/6PHNOM PENHPHNOM PENH, AFP), AFP)

• Cambodia'sCambodia's royal cows signaled a drought royal cows signaled a drought and poor harvests in an ancient ceremony and poor harvests in an ancient ceremony Saturday to mark the beginning of the Saturday to mark the beginning of the kingdom's planting season.kingdom's planting season.

Recent Draft Drought OutlookRecent Draft Drought Outlook

The Three Pillars of the ForecastThe Three Pillars of the Forecast

• Two-week soil moisture forecastsTwo-week soil moisture forecasts

• Climatology: Palmer probabilities, Climatology: Palmer probabilities, analogues, compositesanalogues, composites

• Seasonal temperature, precipitation, and Seasonal temperature, precipitation, and soil moisture forecastssoil moisture forecasts

Features of the Seasonal Drought Features of the Seasonal Drought OutlookOutlook

• A schematic designed to provide a broad indication of A schematic designed to provide a broad indication of drought trends for the following ~3 monthsdrought trends for the following ~3 months

• Four categories: improve, some improvement (e.g., better Four categories: improve, some improvement (e.g., better soil moisture and grassland growth but continuing water soil moisture and grassland growth but continuing water supply shortages), persist/worsen, developmentsupply shortages), persist/worsen, development

• Issued 3Issued 3rdrd and 1 and 1stst Thursday each month (NEW!) Thursday each month (NEW!)

• Incorporates forecasts from all time periodsIncorporates forecasts from all time periods

• Starts with the most recent U.S. Drought monitor D1 areasStarts with the most recent U.S. Drought monitor D1 areas

2-Wk Soil Moisture

Constructed Analogue Soil Model

Medium-Range Fcst

Palmer 4-moProbabilities

CPC Long-LeadPrecip. Outlook

Principal Drought Outlook InputsPrincipal Drought Outlook Inputs

Selected Forecast ToolsSelected Forecast Tools

““These are some of my favorite things.”These are some of my favorite things.”

CPC 2-Week Soil Moisture CPC 2-Week Soil Moisture ForecastForecast

http://www.cpc.noaa.gov/products/soilmst/mrf.shtmlhttp://www.cpc.noaa.gov/products/soilmst/mrf.shtml

Constructed Analogues from Soil Constructed Analogues from Soil (CAS) Moisture(CAS) Moisture

http://www.cpc.noaa.gov/products/soilmst/forecasts.shtmlhttp://www.cpc.noaa.gov/products/soilmst/forecasts.shtml

CFS Modeled Soil MoistureCFS Modeled Soil Moisture

http://www.cpc.ncep.noaa.gov/products/people/wwang/cfs_fcst/http://www.cpc.ncep.noaa.gov/products/people/wwang/cfs_fcst/

Palmer Probability DataPalmer Probability Data

September PalmerSeptember Palmer

Probabilities Probabilities

http://www.ncdc.noaa.gov/oa/climate/research/drought/current.html

Historical Analogues: Soil MoistureHistorical Analogues: Soil MoistureJune 1June 1 Oct. 1Oct. 1

20002000

19871987

http://www.hydro.washington.edu/forecast/monitor/index.shtmlhttp://www.hydro.washington.edu/forecast/monitor/index.shtml

Historical AnaloguesHistorical AnaloguesJune 1June 1 Oct. 1Oct. 1

19861986

19851985

Drought Composites Jun-Sep: PalmersDrought Composites Jun-Sep: PalmersRainfallRainfall TemperaturesTemperatures

June PDIJune PDISep PDISep PDI

Drought Outlook Verification Score

0

10

20

30

40

50

60

70

80

90

100

Jul '03 Oct'03 Jan'04

Apr '04 Jul '04 Oct'04

Jan'05

Apr '05 Jul '05 Oct'05

Jan'06

Apr '06 Jul '06 Oct'06

Jan'07

Date of Forecast

Per

cen

t o

f A

rea

Co

rrec

t

Percent of Area Correct Persistence Score MEAN SCORE MEAN Persistence

Mean score=49%Mean score=49%

Mean improvementMean improvement

over persistenceover persistence

=13%=13%

Verification: Percent Area CorrectVerification: Percent Area Correct

Two Path Drought Forecast ApproachTwo Path Drought Forecast Approach

• Continue current schematic for the Continue current schematic for the general publicgeneral public

• Produce objective probability maps for Produce objective probability maps for technical userstechnical users

Near-term Changes to Drought Near-term Changes to Drought OutlookOutlook

• Have increased frequency to twice/monthHave increased frequency to twice/month

• Propose adjustment of forecast categories Propose adjustment of forecast categories so all categories can be verified (include so all categories can be verified (include category for intensification?)category for intensification?)

• Propose using more current GFS Propose using more current GFS ensemble runs for soil moisture forecasts ensemble runs for soil moisture forecasts and initializing CAS with 2-week forecastand initializing CAS with 2-week forecast

• Considering drought probability mapConsidering drought probability map

Proposed Category ChangeProposed Category Change

Possible Probability MapPossible Probability Map

Drought Recovery Prediction MethodologyDrought Recovery Prediction Methodology

• Use observed forcings to drive the model and produce “best knowledge” Use observed forcings to drive the model and produce “best knowledge” initial conditions for forecastinitial conditions for forecast

• Drive the model with precipitation and temperature forcings unconditionally Drive the model with precipitation and temperature forcings unconditionally sampled from climatology creating an ensemble of model trajectoriessampled from climatology creating an ensemble of model trajectories

• At different lead times, estimate probability of soil moisture/runoff At different lead times, estimate probability of soil moisture/runoff percentile exceeding threshold from ensemble = probability of recoverypercentile exceeding threshold from ensemble = probability of recovery

Prototype forecast product from UWPrototype forecast product from UW

Probability of Drought RecoveryProbability of Drought RecoveryInitial Conditions (2/2006) 1-month lead forecast (3/2006)

6-month lead forecast (8/2006)3-month lead forecast (5/2006)

Prototype forecast product from UWPrototype forecast product from UW

Current Princeton Probability Current Princeton Probability ForecastsForecasts

Thoughts for the FutureThoughts for the Future

• Statistical and dynamic (GCM) techniques can be Statistical and dynamic (GCM) techniques can be used to forecast the probabilities for reaching used to forecast the probabilities for reaching various levels of the drought variable—need to various levels of the drought variable—need to effectively combine ST and LT forecastseffectively combine ST and LT forecasts

• For a drought early warning system, need to For a drought early warning system, need to consider current conditions, trends, and consider current conditions, trends, and forecastsforecasts

• Effective early warning systems (DEWS) provide Effective early warning systems (DEWS) provide alerts (e.g. watch, warning) and assessment, not alerts (e.g. watch, warning) and assessment, not just data. “Alert” for developing droughts? just data. “Alert” for developing droughts? “Advisories” for ongoing droughts? “Advisories” for ongoing droughts?

Do we ringDo we ring

a bell fora bell for

drought?drought?