ODFW's Oregon Plan Habitat Monitoring Program
for Coastal Basins

  Home Landowner Info Habitat Data Habitat Data Precision Habitat Trends

Precision: ODFW Habitat Resurvey and Analysis

An analysis of survey precision was completed for the 2000 dataset and for the combined 1998-2000 data.  The precision of an individual survey metrics was calculated using the mean variance of the resurveyed stream’s “Noise” and the overall variance encountered in the habitat survey’s “Signal”.  Three measures of precision were calculated: the standard deviation of the repeat surveys (SDrep), the coefficient of variation of the repeat surveys (CVrep), and the Signal to Noise ratio (S:N).  S:N ratios of less than 2 can lead to distorted estimates of distributions and limit regression and correlation analysis.  S:N ratios between 2 and 10 are useful for analysis, but caution must be exercised due to the larger variances associated with each variable.  S:N ratios greater than 10 are very good and indicate that variables have insignificant error caused by field measurements and short term habitat fluctuations (Kauffman et al. 1999).

An assessment of variable precision incorporated annual and composite variability for 1998-2000 (Table 9).  The complete dataset encompassed 724 habitat survey sites and 85 sites where a repeat survey was conducted.   The signal to noise ratio for channel length, gradient, percent dammed pools, deep pools per km, and 0.3 m conifers per 305 m of stream length varied widely between years, while the precision (reported as the standard deviation and coefficient of variation) remained consistently low.  Measurements of signal to noise between years were consistent for percent secondary channels, percent fines, percent gravel, percent fines in riffles and percent gravel in riffles.   Signal to noise ratios for woody debris were low but precision remained high.  As in previous years, it appeared that a few sites with large amounts of wood proved difficult to count consistently thereby reducing the overall accuracy of the wood data collected.

            All signal to noise ratios for variables in the combined 1998-2000 dataset exceeded a value of 2 with several dependent and independent variables with high ratios.  Reliable independent variables included channel length, channel width, floodprone width and gradient.  Reliable dependent variables were percent secondary channels, percent pools, percent dammed pools, deep pools per km, percent fines, percent bedrock and percent fines in riffles.

 Table 9: Signal to Noise Ratios for 2000 and combined 1998-2000

Independent Variables:

Variable Year S.D. (repeats) CV S:N*

Channel Length

 1998
1999
2000
1998-2000

47.8
26.7
24.9
34.8

 6.6
3.5
3.5
4.8

29.8
93.8
104.2
55.2
Channel Width 1998
1999
2000
1998-2000		 1.3
1.7
1.5
1.6		 18.1
19.5
17.4
18.4		 13.7
29.8
26.4
27.8
Floodprone Width 1998
1999
2000
1998-2000		 3.7
3.4
3.3
4.4		 25.9
27.6
22.0
32.0		 10.0
11.2
36.1
10.4
Gradient 1998
1999
2000
1998-2000		 0.5
1.8
0.8
1.1		 8.9
31.6
16.2
20.7		 172.9
11.8
51.4
30.9

*S:N = Variance of all surveys/variance between repeat surveys

Dependent Variables:

Variable Year S.D. (repeats) CV S:N*
% Secondary Channels 1998
1999
2000
1998-2000

3.0
3.1
3.0
2.9

 70.0
66.2
74.2
68.9

4.3
7.9
4.5
5.6
% Pools 1998
1999
2000
1998-2000

8.1
7.7
6.0
7.1

 30.2
23.7
17.5
23.2

6.7
27.3
17.2
17.0
% Dammed Pools 1998
1999
2000
1998-2000

0.9
5.6
2.8
3.9

 18.5
112.2
48.3
76.6

235.7
6.8
30.5
13.6
Deep Pools/km 1998
1999
2000
1998-2000

0.7
1.1
1.0
1.2

 28.9
54.1
46.1
51.5

33.4
5.8
11.4
9.3
Residual Pool Depth 1998
1999
2000
1998-2000

0.3
0.1
0.1
0.2

 55.7
13.5
12.2
31.9

1.7
14.4
14.1
3.5
% Sand/Organics 1998
1999
2000
1998-2000

6.8
7.8
5.8
6.7

 23.6
26.2
19.5
22.9

11.5
9.1
18.0
12.4
% Gravel 1998
1999
2000
1998-2000

9.1
7.6
8.4
8.1

 36.3
27.6
28.3
30.0

2.0
4.1
3.5
3.2
% Bedrock 1998
1999
2000
1998-2000

2.9
2.8
4.6
3.5

 27.1
27.9
47.8
34.4

21.6
20.3
8.2
13.9
% Riffle Sand/Organic 1998
1999
2000
1998-2000

7.6
7.6
10.2
8.6

 30.2
29.7
41.5
34.1

7.6
8.7
5.4
6.7

% Riffle Gravel

 1998
1999
2000
1998-2000

9.5
10.3
16.1
11.9

 28.3
26.2
38.9
31.6

3.3
4.5
1.6
2.8

Wood Pieces / 100m

 1998
1999
2000
1998-2000

3.6
4.2
10.7
9.4

 24.9
23.8
73.4
60.8

13.4
2.1
2.0
2.2

Wood Volume / 100m

 1998
1999
2000
1998-2000

7.4
9.4
17.5
15.8

 34.2
35.7
64.5
64.3

11.0
2.5
3.9
3.6

Key Wood Pieces / 100m

 1998
1999
2000
1998-2000

0.6
1.5
1.1
1.1

 70.9
136.5
91.5
108.4

3.8
1.7
3.5
2.4

Wood Jams / km

 1998
1999
2000
1998-2000

2.6
1.7
2.3
1.9

 52.4
36.6
60.1
49.7

5.3
6.4
3.3
4.5

Shade

 1998
1999
2000
1998-2000

5.2
6.2
10.3
7.6

 6.7
7.5
13.5
9.6

11.5
6.2
3.7
5.5

0.5 m Conifers / 305 m

 1998
1999
2000
1998-2000

20.0
69.4
19.3
42.4

 49.5
98.3
54.0
88.3

10.0
2.3
7.3
3.3

1.0 m Conifers / 305 m

 1998
1999
2000
1998-2000

6.0
32.8
4.6
23.1

 59.0
161.7
125.6
168.3

24.6
1.6
4.0
2.1

*S:N = Variance of all surveys/variance between repeat surveys

Back to ODFW Oregon Plan Habitat Monitoring

Back to ODFW Aquatic Inventories Project