Ecological Research

1
Ecological Research LTG 1 Poster 1
EMAP Monitoring Design Design Team Anthony
(Tony) R. Olsen (USEPA), N. Scott Urquhart
(Colorado State U), Don L. Stevens (Oregon
State U)

• 305(b) Status Trends
• More efficient survey designs
• Better statistical analyses

Small area estimation Making available data do
more

• Statistical Research
• Collaboration among ORD researchers and STAR
  Grant statistical researchers
• Over 250 peer-reviewed publications
• Invited monitoring program reviews (e.g., NOAA
  Mussel Watch, Pacific Rim Salmon monitoring,
  Everglades restoration, Grand Canyon, Alberta
  biodiversity, NPS inventory monitoring)
• Conferences organized
• Computational Environmetrics 2004
• Monitoring Science Technology Symposium
  Statistical track, 2004
• Graybill Conference on Spatial Statistics
• 4 Fellows American Statistical Association

Use EMAP probability survey data from 557 lakes
to estimate average lake ANC for 113 Hydrologic
units. Requires auxiliary data and new
semi-parametric statistical methods
GRTS Spatially-balanced sampling Improvement
over simple random or systematic sampling
Predict likelihood of water-quality impaired
stream reaches from probability survey and
auxiliary data, e.g., landscape characteristics
relevant to 303(d)

• EMAP Design Team
• Members from 4 NHEERL Eco-divisions, 2 NERL
  divisions, Office of Water and EPA Regions
• Mechanism to transfer statistical research to
  EPA and state monitoring designs while team works
  with states
• Technical Transfer
• Aquatic Resource Monitoring website
  \\www.epa.gov\nheerl\arm
• Software for site selection and statistical
  analysis psurvey.design psurvey.analysis
• Monitoring workshops for states and EPA Regions
  (over 10)
• Internet meeting training sessions with
  individual states on monitoring design analysis
• 30-40 monitoring designs per year for states,
  EPA, and other federal agencies (USGS, NPS,
  NMFS, USFS)

Improved variance estimation Better precision
for fixed cost.
Develop methodology using Maryland Biological
Stream Survey data

• A geostatistical model
• Predict a specific reach scale condition at
  points that were not sampled
• Provide a better understanding of the
  relationship between the landscape and reach
  scale conditions
• Give insight into potential sources of water
  quality degradation
• Develop landscape indicators
• Crucial for the rapid and cost efficient
  monitoring of large areas
• Better understanding of spatial autocorrelation
  in stream networks
• What is the distance within which it occurs?
• How does that differ between chemical variables?
• 3. Produce map of study area
• Shows the likelihood of water quality impairment
  for each stream segment
• Based on water quality standards or relative
  condition (low, medium, high)
• Future sampling efforts can be concentrated in
  areas with higher probability of impairment
• 4. Transfer technology to States and Tribes

Relative Risk Estimation The risk of Poor BMI
is 1.6 times greater in streams with Poor SED
than in streams with OK SED.