Development of Aquatic Ecosystem Models

1
Development of Aquatic Ecosystem Models

• Lizhu Wang, Shaw Lacy, Paul Seebach,
• Mike Wiley
• Institute for Fisheries Research
• MDNR and U of M

2
Project Objectives

• Develop statewide models for predicting stream
  flows and summer water temperature
• Develop models describing relationships among
  base flow, water temperature, and trout
  population
• Test model sensitivity by incorporating
  site-specific data from Augusta Creek
• Evaluate effectiveness of catchment BMPs in
  offsetting ground water withdraw.

3
The Key Step Describe relationships among flow,
temperature, and trout
TROUT
FLOW
TEMPERATURE
4
Build Statewide GIS Databases

• Identify stream measuring unit

5
Michigan
1100,000 NHD
38,000 inter-confluence stream reaches
6
Build Statewide GIS Databases

• Delineate catchment boundaries for each unit

7
Reaches, Watersheds, Riparian Zones, and Upstream
Catchments
8
Upstream Catchment Context
9
Build Statewide GIS Databases

• Synthesize landscape instream characteristics
  for each stream unit

10
(No Transcript)
11
Synthesize Landscape Data

• Land use/land cover
• Surficial geology texture formation
• Soil permeability
• Bedrock depth geology
• Average annual growing degree days
• Average annual precipitation
• Air temperature
• Ground water delivery potential Darcy
• Land use transformation model output

12
Synthesize Network Instream Data

• Arc gradient
• Catchment gradient
• Arc sinuosity
• Arc stream order, linkage number
• Arc fragmentation by dams and linkages with
  lakes and large rivers

13
Variables and Scales

• The database contain about 300 variables.
• The data are organized into 5 scales
• Channel
• Riparian (arc)
• Riparian (entire)
• Watershed (arc)
• Watershed (entire)

From Paul Steen, 2004
14
Build Sampled Databases

• Link sampled flow, temperature, and trout data
  with each stream unit

15
Develop Summer Water Temperature Model

• Collected continuous summer water temperature
  from about 500 sites throughout Michigan.
• Developed geo-statistical and generalized
  additive regression models for July mean ----
  Temperature kriging residual modeling (Mallows
  Cp, Min AIC).
• Model explains 77 variance.

16
Michigan Stream Temp. Predictions
26.0º C
9.0º C
17
Flow Models
Based on 83 sites that have gt20 years flow data

• Multiple regression models for predicting
  annual and August 5, 10, 25, 50, 75, 90,
  95 exceeding flows
• Models explain 78 97 variance.

18
Annual Flow Predictions 90 Flow Yield
19
Stream Sites Sampled for Fish Community

Based on fish data from about 1500 sites
20
Link flow, Temperature, and Trout (hypothetical
model)
Cold
Typical
Marginal
July mean water temperature oC
21
Cold
Typical
Marginal
22
Link flow, Temperature, and Trout (hypothetical
model)
Cold
Typical
Marginal
July mean water temperature oC
23
Cold
Typical
Marginal
24
(No Transcript)
25
(No Transcript)
26
Summary

• Identified stream measuring unit
• Delineated catchment boundaries of each unit
• Synthesized landscape instream characteristics
  of each unit
• Linked sampled data for trout, flow, and
  tempertaure with each stream unit.
• Developed models for prediction of flow and
  temperature.
• Partially built models linking trout, flow, and
  temperatur.
• Collected temperature and fish data from Augusta
  Creek.

27
(No Transcript)
28
(No Transcript)