Transcript Ecological Nowcasting in Chesapeake Bay Christopher Brown NOAA Satellite Climate Studies Branch

Ecological Nowcasting in
Chesapeake Bay
Christopher Brown
NOAA Satellite Climate Studies Branch
CICS - ESSIC
University of Maryland, College Park
Importance of Coastal Ocean
Monitoring & Prediction
• National Goal
– Congress initiated efforts to establish a coastal monitoring system
and develop coastal hydrodynamic models
• NOAA Goal
– VADM Lautenbacher stated that an ecosystem assessment and
prediction capability was a critical NOAA to provide information
on coastal and marine ecosystems
– He also wrote that by 2011 NOAA “should be able to forecast
routinely the extent and impact of critical ecosystem events, such
as harmful algal blooms”
• Biological Oceanography Goal
– Develop the understanding and the means to detect and predict
distribution pattern of organisms
Motivation for Study
• Detect and predict
distribution pattern
of organisms that
affect society, both
beneficial and
harmful
• Few existing
methods work well
and in near-real
time
Bloom of the coccolithophorid Emiliania huxleyi
in the Barents Sea in July 2003 in SeaWiFS
imagery. Image courtesy of NASA SeaWiFS Project and OrbImage.
Approaches for Predicting Organisms
• Process-Oriented or Mechanistic
Modeling
• Empirical or Statistical Modeling
Mechanistic Modeling
Statistical Modeling
• Develop multi-variate empirical habitat
models
– Quantitatively define the preferred
environmental conditions of the organism
• Based on Concept of Ecological Niche
– Identify the geographic locations where
ambient conditions coincide with the
preferred habitat of target organism
Hybrid Statistical – Mechanistic Approach
• Develop multivariate empirical
habitat models
• Drive habitat
models using
real-time data
acquired from a
variety of sources
Habitat Model
Hybrid Statistical – Mechanistic
Ecological Approach
• Old technique employed in new way
– GAP Analysis: retrospective analysis
– Ecological Nowcasting: near-real
time
Ecological Nowcasting In Chesapeake Bay
Currently generate
nowcasts of two
species in
Chesapeake Bay
• Sea Nettles,
Chrysaora
quinquecirrha
• Dinoflagellate
Karlodinium
micrum
Chance of encountering sea
nettle, C. quinquecirrha, on
August 15, 2004
Relative abundance of the
harmful algal bloom K. micrum
on May 27, 2004
Nowcasting Sea Nettle Distributions
in Chesapeake Bay: An Overview
C. W. Brown1, R. R. Hood2, T. Gross3,
Z. Li3, M.-B. Decker2, J. Purcell2 and
H. Wang4
1NOAA/NESDIS
Office of Research & Applications
2Horn Point Laboratory, UMCES
3NOAA/NOS Coast Survey Development Laboratory
4VIMS, College of William and Mary
Funded by NORS Grant, Maryland SeaGrant, NCCOS EcoFore 04
Chrysaora quinquecirrha
(Photo by Rob Condon)
Introduction: Sea Nettles
• Chrysaora ephyra and
medusa seasonally
populate Chesapeake
Bay
ephyra
• Chrysaora is
biologically important
and impacts recreational
activities
• Knowing the distribution
of Chrysaora would
provide valuable
information
juvenile
medusa (adult)
egg
strobila
scyphistoma
polyp
larva
Life Cycle of Chrysaora
From: T.L. Bryant and J.R. Pennock (eds). 1988. The Delaware Estuary: Rediscovering a
Forgotten Resource. University of Delaware Sea Grant College Program. Newark, DE.
Sea Nettle Nowcasting Procedure
1. Estimate current
surface salinity and
temperature fields
SST
Likelihood of Chrysaora
2. Georeference salinity
and SST fields
3. Apply habitat model
4. Generate image
illustrating the
likelihood of
encounter of
Chrysaora
Habitat Model
Salinity
Surface Salinity
35
• Generated using
hydrodynamic model
developed for the Chesapeake
Bay
30
25
• Model forced using near-real
time input
20
15
• Model attributes:
– Horizontal Resolution: 1-5
kilometers
– Vertical Resolution: 1.52
meters
– Error: 2 - 3 ppt
10
5
0
Model generated surface salinity in
Chesapeake Bay for April 20, 2005
Sea-Surface Temperature
35
Two Sources:
– Error: 2 - 3 °C
2. Derived from NOAA
AVHRR satellite imagery
– Resolution: 1 km
– Weekly composite
– Bias: 0.5 °C; STD: 1.0°C
25
20
15
10
5
0
Model generated sea-surface temperature
in Chesapeake Bay for April 20, 2005
Sea-surface Temperature (ºC)
1. Generated by
hydrodynamic model
30
Sea Nettle Habitat Model
• Models developed to predict:
39.5
1. Probability of encountering
Chrysaora
2. Density of Chrysaora
• Samples collected in surface
waters (0 –10 m) of
Chesapeake Bay (n = 1064)
– 2/3 model training
– 1/3 model testing
Latitude ( N)
• Analyzed relationship
between Chrysaora, salinity
and sea-surface temperature
39.0
38.5
38.0
37.5
37.0
-77.0
-76.5
-76.0
Longitude ( W)
-75.5
Sea Nettle Habitat
% Medusa presence
80
60
40
20
0
0
5
10
15
20
Temperature (Co)
25
30
0
5
10
15
20
Salinity (difference from optimum)
Nettle medusa occupy narrow temperature (26-31 °C) and
salinity (10-16 PSU) range. Salinity optimum = 13.5
PSU.
25
Probability of Encountering Sea Nettles
• Combination of salinity
and SST is a good
predictor of Chrysaora
presence
• If SST < 34°C:
– p = elogit / (elogit + 1),
where,
logit = -8.120 + (0.351*SST)
- (0.572* |SAL - 13.5|)
– Hosmer-Lemeshow Goodness
of Fit P = 0.493
Probability of medusa occurrence
1.0
0.8
0.6
0.4
0.2
0.0
0
Absent
1
Present
2
Observed medusa occurrence
3
Nowcasting the Relative Abundance of
Karlodinium micrum in Chesapeake Bay
Christopher W. Brown1, Douglas L. Ramers2,
Thomas F. Gross3, Raleigh R. Hood4, Peter J.
Tango5 and Bruce D. Michael5
1NOAA, 2University
Of Evansville, 3NOAA & Chesapeake Research Consortium,
4University of Maryland Center for Environmental Science – Horn Point Laboratory,
5Maryland Department of Natural Resources
Project Funded by NOS MERHAB Program
Karlodinium micrum
 A common estuarine
dinoflagellate found along
the U.S. East Coast
 Seasonally abundant in
Chesapeake Bay
 Contributed to several fish
kills in Chesapeake Bay
 Significant blooms
confined to a relatively
narrow range of salinity
and temperature
Photomicrograph of the dinoflagellate
Karlodinium micrum.
K. micrum Nowcasting Procedure
1. Estimate current
surface salinity and
temperature fields
SST
Relative Abundance of
K. micrum
2. Georeference salinity
and SST fields
3. Apply habitat model
4. Generate image
illustrating the
relative abundance of
K. micrum
Salinity
Habitat Model
Habitat Model
o Neural Network (NN) employs sea surface
temperature, salinity and month to predict the
relative abundance of K. micrum at low, medium and
high or “bloom” concentrations
o NN trained with samples (n = 151) of in-situ K.
micrum abundance and various environmental
variables
o A test data set (n = 81) was extracted from the
available data to assess the model’s performance
Schematic Representation of
Neural Network
Input Layer
X1
I
N
P
U
T
S
X2
X
X3
h
h
h
Xn
wij
Hidden Layer
Output Layer
PE1
f (W h X + b)
aPE
PE2
f (W h X + b)
h
h
h
PEm
f (W h X + b)
Xi * wij
aPE
aPE
PE out
f (W h X + b)
Classify
a=
1
0
-1
= foutput(Woutputhfhidden (Whidden h X + bhidden) + boutput)
Issues and Advantages of Neural Networks
• Issues
– “Black Box”
• Advantages & Uses
– Useful for representing and processing inexact
and sparse data and for performing
approximate reasoning over uncertain
knowledge and ill-defined problems
– Useful in discerning patterns and relationships
– No a-priori distribution assumed
K. micrum Neural Network Performance
Count Confusion Matrix:
Predicted
Low
Med.
High
Low
26
4
2
32
Actual
Med.
5
11
2
18
High
0
0
31
31
31
15
35
81
Frequency Confusion Matrix:
Actual
Low
Med.
Low 81%
28%
Predicted
61%
Med. 13%
High
6%
11%
High
0%
0%
100%
Nowcast vs. In-Situ Comparison
May 27, 2004 - Nowcast
0-10 cells/ml
10-2000 cells/ml
>2000 cells/ml
May 23-26, 2004 - In-situ
Nowcast WWW Sites
Sea Nettle and K.
micrum nowcasts
are generated daily
and are available
on the World Wide
Web.
http://coastwatch.noaa.gov/seanettles
http://coastwatch.noaa.gov/cbay_hab/index.html
Future Directions and Work
o Continue nowcast validation and refine habitat
models of Chrysaora and Karlodinium
o Develop habitat models for additional HAB species
in Chesapeake Bay
o Incorporate additional environmental variables into
habitat models and nowcast system to enhance HAB
prediction capability
o Generate historical distribution patterns of occurrence
and relative abundance from retrospective salinity
and temperature to document interannual variability
Issues With Empirical Approach
• Empirical models are specific for each
location and population
• Development of empirical models require
sufficient number of samples
• Species acclimate to environment, i.e.
habitat model may change
Regional Ecosystem Modeling
• Objective: Develop a fully
integrated, bio-physical model of
Chesapeake Bay and its watershed
that assimilates in-situ and
satellite-derived data.
• Purpose:
– Near-Real Time Applications:
Nowcasting and forecasting of
marine organisms, ocean
health, and coastal conditions
– Climate Research: Estimating
effect of climate change on the
health of coastal marine
ecosystems
• Partners: NOAA, CICSESSIC, other UMD departments,
Meteorology, and programs, e.g.
UMCES.
SeaWiFS True-Color Image of Mid-Atlantic Region
from April 12, 1998.
Image provided by the SeaWiFS Project, NASA/Goddard Space Flight Center and ORBIMAGE
Regional Ecosystem Model
Plans & Objectives
• Develop transportable modeling system that
can be modified for other regions
– Chesapeake Bay used as “test bed” site due to
extensive in-situ data for verification
• Employ satellite imagery in system for
monitoring, model forcing and data
assimilation to permit use in locations where
in-situ assets are limited
Advanced Study Institute for
Environmental Prediction
• Institute dedicated to research on environmental
prediction and monitoring
– Perform research and provide core support to determine
what present and future observations need to be
sustained beyond numerical weather prediction in
support of Earth system predictive models, crops
models, and predictive disease models
• Staffed by personnel from NOAA, NASA
Goddard, and the University of Maryland
• $1.5M budgeted for Institute in FY06 2006
Science, State, Justice and Commerce
Appropriations conference report
Thank You!
10
100
a
Observed
b
80
Predicted
8
Forced
60
6
40
4
20
2
0
0
30
28
26
24
22
20
18
16
14
c
d
20
e
f
Salinity
18
16
14
12
10
May
Jun
Jul
Aug
Date
Sep
Oct
Nov
Jun
Jul
Aug
Date
Sep
Oct
Nov
Predicted likelihood
of occurrence (%)
Observed medusa abundance
Temperature (oC)
12
Interannual Variability
Probability of Encountering C. quinquecirrha
July 25, 1996
July 29, 1999
Likelihood of Encountering C. quinquecirrha
in July 1996 and 1999
Vibrio cholerae
• Presence predicted
as function of
water temperature
and salinity (Louis
et al., 2003)
• Association with
plankton
Electron photomicrograph of Vibrio cholerae:
curved rods with polar flagellum.
http://microvet.arizona.edu/Courses/MIC420/lecture_notes/vibrio/em.html