NOS Center for Coastal Monitoring and Assessment

HAB forecasting C NI O D ATM SPHER AN IC TRATION NIS MI AD NATIONAL OC EA Shelly Tomlinson NOAA/NOS Center for Coastal Monitoring and Assessment ...
Author: Claribel Black
1 downloads 0 Views 1MB Size
HAB forecasting

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Shelly Tomlinson NOAA/NOS Center for Coastal Monitoring and Assessment

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

What are Forecasts? •

Nowcast: There is a HAB, where is it today? ƒ This is usually where it was yesterday. A location (detection/extent) and a transport problem



Forecast: There is a HAB, where will it be in the next few days? What will be the likely impact? ƒ A location (detection/extent) and a transport modeling problem *ONLY AS GOOD AS BLOOM DELINEATION ESTIMATES



Prediction:

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

ƒ What kind of HAB ƒ When and where will HAB initiate or dissipate? Will the HAB still be here next month? ƒ How severe? No HAB today, but HAB next week/month? ƒ These are ecological modeling problems

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

What is operational? • Scheduled ƒ During active bloom, comprehensive analysis and graphics e-mailed at least twice a week (subscription only) ƒ Public information updated twice-weekly ƒ Web-based ƒ Current conditions and expected conditions for next 1-3 days ƒ Bulletins available to the public after minimum 1 week delay

• Reliable ƒ Operational on a 5-day week (analysts available every day). ƒ Analysts available at [email protected] and by cell phone to bulletin subscribers

• Institutional

C NI

D

ER

S.

CE

U. EP

NOAA Operational Office Standard operational procedures User training Annual assessment

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

ƒ ƒ ƒ ƒ

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

NOAA HAB Forecast System Operational since 2004

Observations

Model output

(satellite imagery, buoys, field samples, respiratory irritation)

(heuristic, empirical)

Synthesis and Analysis

A harmful algal bloom has been identified in patches from southern Lee to central Collier County. Patchy very low impacts are possible from southern Lee County to central Collier County today through Thursday. No other impacts are expected.

Conditions Report (public)

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

HAB Bulletin (managers) D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

HAB-Operational Forecast System a Success http://tidesandcurrents.noaa.gov/hab/



HAB-OFS has been in place for over 4 years



NOAA has issued nearly 400 bulletins in that time in response to HABs that have impacted over 1000 km of coast



90% weekly utilization* by over 190 resource managers representing more than 50 organizations



Integration with Public Health (e.g., Collier County Pollution Control and Prevention, Florida Aquatic Toxins Hotline, Florida DOH, CDC)

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

*Usage based on reference in weekly reports (state or county) Samples in blooms D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Rule-based models for analysis (Scientific SOP) •



C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA



D

ER

S.

CE

U. EP

Established rules for development of a bloom ƒ Timing of blooms (seasonality) ƒ Location of blooms (identify hot spots, areas where blooms do not occur) Once established at the coast, a series of rules for forecasts ƒ Transport: where is bloom going ƒ Extent: will bloom expand its region of impact ƒ Intensification: will bloom get bigger (higher cell counts/larger area) ƒ Impact: what conditions will beaches experience, e.g., fish kills, respiratory irritation Needed improvements: ƒ Conditions for bloom initiation and demise ƒ Sea breeze model for respiratory impacts

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Evaluation of forecasts important part of operational system From Oct 2004-Apr 2006 ƒ Accuracy was high because of low resolution forecasts ƒ Efforts underway to improve resolution

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Stumpf et al., 2009 Journal of Marine Systems

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Public Advisory – new with operational system

Current conditions and forecast of beach impacts over the next few days

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Precision/Accuracy only as good as ability to identify bloom field

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

FL Beach Impact (Lifeguard)

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Annual Skill Assessment • Provides better understanding of respiratory irritation forecasts County-wide Forecast of moderate/high impact Correct County- 73% wide (at least one beach) Correct against 21% individual reports from 6 beaches

C NI

D

ER

S.

CE

U. EP

Stumpf et al., 2008, J.Marine Systems

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

coolgate.mote.org/beachconditions AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Importance of Resolution on Forecast Accuracy Resolution at the Coast (Florida) Consistent with the initial HAB “field”

30 km per day

Forecast Resolution

30-60 km per day

Accuracy at 30-60 km

73% correct

Accuracy at 10 km

21% correct

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Resolution is only as good as the resolution of the initialization data. (1 km satellite pixels do not produce 1 km/day resolution in the HAB Field) D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Efforts to fill data gaps • Take advantage of public health ƒ ƒ ƒ ƒ ƒ

Poison control hotline Beach reports from lifeguards FWRI fish kill database (for respiratory irritation reports) Turtle Patrol reports HABISS

• Addition of brevebuster on NOAA NWLON Naples Pier, FL • Brevebusters to be added to NOAA CMAN (3 in Veracruz; Venice, FL). • Purchase of two gliders under research $$

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

ƒ However, funds unavailable to run operationally

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Assimilation strategy for analysis in Google Maps

Example analyst page for Florida cell counts lifeguard obs, fixed sensors

glider (not operational)

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

satellite D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

C NI

D

ER

S.

CE

U. EP

Data Type

Source

Format

Cell Counts (FL, AL, MS)

Direct, also FWRI database

Manual, Excel, PDF (parsed), DBF

Cell Counts (TX)

TexHAB (integrated by TPWD)

Manual, Excel

Alerts/Warnings of Problems

Direct

Email lists and direct email

Respiratory Impact Data (Lifeguard program)

Mote Marine Lab

CSV, ftp/email (automated process)

AUV/Moored brevebuster detection

Mote Marine

ftp of csv (automated process)

Poison Hotline

U.Miami

Automated ftp

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Data sources critical to identifying a HAB field: needs for standardized formats/data delivery

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Texas: Demonstration since 2006

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Expanded monitoring in TX

Volunteer Phytoplankton Monitoring Network • Established by NOAA’s Marine Biotoxins Program, 2001 in SC • Unites volunteers and NOAA scientists • Sampling takes place on a weekly or bimonthly basis • Equipment provided • Results enable researchers to identify problem areas

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

•Expanded to TX in 2007 D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Texas: A coordinated success story

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Texas, TexHAB averted a disaster: March 2008: Dinophysis in Corpus Christi Bay, never reported before as a HAB in the GOM First spotted by TAMU experimental video sensor Linkage across Texas community through TexHAB Support by various groups: sampling by PMN, modeling support by NOAA and TGLO/TAMU. Week of major oyster festival. Oyster recall; fresh oysters for festival; no cases of DSP. D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Other U.S. Gulf States

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

• Ad hoc event response when unusual event occurs • No system for LA/MS/AL • If event originates in FL, the operational system will continue to support

D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Need for US/Mexico communication Dead fish wash up on Nassau beaches (Florida Times Union, 25/9/2007) Thousands of dead fish have washed up on Nassau County beaches, and beachgoers have reported scratchy throats, irritated eyes and coughs while exposed to the air and water.

SeaWiFS Chl (13/9/2007)

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

Naval Research Laboratory HYCOM Model Output (15/9/2007) D

ER

S.

CE

U. EP

AR

TME

O NT OF C

M

M

HABIOS Workshop (St. Petersburg, FL)

21 April 2009

Forecast needs from HABIOS • Priority: Location/Validation data at appropriate (14 km/d) resolution

C NI

O D ATM SPHER AN IC

TRATION NIS MI AD

NATIONAL OC EA

ƒ HAB field currently based on satellite/cell counts (1km satellite, obscured by clouds, 50 cells counts/week for 200 km of coastline) ƒ Respiratory irritation data (increased errors to

Suggest Documents