John's Creek Watershed Analysis March 19, 1999 University of Washington Forest Engineering 423/523

John's Creek Project Team

Introduction

Links:

• Riparian Module
• Channel Code
• Riparian Code

Channel Module

Overview

• Spatial distribution of channel response types
• Evidence of change from historic condition
• Indications of past and present active geomorphic processes
• Response to changes in input factors
• Channel and habitat forming processes

Not all of this questions can be addressed through a computer environment, some of them are intended for direct field inspection. The spatial distribution of channel response type and response to changes in LWD were answered and the results used as inputs for other modules (riparian, fish)

Data inputs

The data we had to work with had limitations and errors that influenced the channel module as well the whole analysis.
The stream coverage was delineated as a minimum power (stress) network using formula:

        Stress = depth * slope * unit weight of water      (force / unit area)

The original digitized DNR stream coverage, considered most accurate was the starting point then the digital elevation model was forced to align with that. The streams were then generated from the new DEM with the imminent consequences of precision lost.
 

Method

The core of the analysis is the WAM stream response matrix (table E-2), that classifies geomorphic processes according to stream characteristics:
 

unconfined	fs,be,wa	wl,sf,fs,be	db,dfd,be,cs,sf,wl	dfs,dfd,db,wl	dfs
mod.conf	fs,be,wa	cs,be,sd,wl,fs,	cs,be,db,sd,dfd,wl,sf	dfs,dfd,db,sf,wl	dfs	dfs
confined		cs,wl	cs,sd,wl,dfd,db	dfs,dfd,db,sf,wl	dfs	dfs
	grad < 1	grad 1-2	grad 2-4	grad  4-8	grad 8-20	grad > 20

                                   table 1.Valley gradient and typical channel bed morphology

where fs - fine sediment deposition, cs - coarse sediment deposition, sd- scour depth, sf - scour frequency, be - bank erosion, wl - wood loss, wa - wood  accumulation, dfs - debris flow scour, dfd - debris flow deposition, db - dam break flood

The analysis was performed both in ArcGrid and ArcView, trying to develop an easy to use tool that will give non-GIS user an easy way to solve the problem.
Details of tool construction available with the code.

Due to the lack of stream bed and valley width data stream confinement had to be estimated using surface curvature. Determining confinement class breaks is an important factor in accurate modeling of stream response. Class breaks were equally set at thirds of curvature range, based on class histograms and aerial photos as reference. Such an approach has yet to be proven by field inspection. Most of John's Creek watershed fall in the moderately confined and unconfined categories.

Stream sensitivity to Large Woody Debris inputs is considered under the Riparian Module
 

Results

        Sediment Deposition

        Discharge Sensitive

        Wood Loss / Accumulation

        Catastrophic Events
 

Conclusions

Analyzing the above, John's Creek watershed appears dominated by high gradient streams that facilitate catastrophic events and wood loss processes. Fine sediment deposition hardly occurs in lower areas towards the output point of the basin and discharge sensitive processes are well represented.
 
 

Riparian Module

Overview

Riparian areas are important for supporting environmental requirements of fish from the point of view of both water temperature and presence of wood  in streams. The goal of this module is to provide an understanding of the current condition of riparian areas and impacts of possible logging.
The module is structured in two parts: Riparian Shade and Large Woody Debris Recruitment. For each part a comparison of the actual case with the scenario of harvested 50 year or older stands was considered.

Data inputs

The module requires both spatial and land cover data provided by DNR's RIU and DEM. Grids of dominant species, canopy percent, average tree height and diameter at breast height were generated from RIU stand coverage. Our confidence in data is low due to data inconsistencies observed (especially regarding DBH).

        Canopy Percent

        Diameter at Breast Height

John's Creek is entirely covered by Douglas Fir as primary species- image not shown.
Sun altitude above horizon and Sun azimuth were also necessary. The data was collected from the U.S. Naval Observatory http://aa.usno.navy.mil/AA/data for the City of Shelton, Aug. 1st 1998. NOAA records for a period of 30 years show Aug. 1st as the hottest day of the year.
 
 

Astronomical Applications Dept.  U.S. Naval Observatory  Washington, DC 20392-5420 SHELTON, WASHINGTON     o  ,    o  ,  W123 07, N47 13

Altitude and Azimuth of the Sun  Aug. 1, 1998  Pacific Standard Time

Altitude    Azimuth                       (E of N)  h  m         o           o

04:00       -8.1        52.2 06:00       10.1        74.3 08:00       30.2        95.9 10:00       49.2       124.0 12:00       60.5       170.9 14:00       54.1       223.9 16:00       36.4       256.4 18:00       16.2       279.1 20:00       -3.0       300.5

table 2. Sun altitude and azimuth

Riparian Shade

Method

Riparian shade is estimated as the difference of the available shade and the desired target shade. The target shade is established with the TWF temperature standards for Western Washington; for the study case, non glacier fed class A streams the standard is 18.3 degree Celsius.
Minimum shade is a function of elevation and was derived from the modified DNR DEM.

    Riparian Target Shade

Riparian shading is considered effective for all streams less than 100' wide. Stream width and depth were estimated using Dunne and Leopold's "average values of bank full channel dimensions as functions of drainage area" for San Francisco Bay region (graph A). All of John's Creek streams are below this threshold.
  fig 1. Stream width and depth function of contributing area

        Stream Width

Available shade depends on stream width, tree height, stand density, local topography and Sun's position on a particular moment of the day. For this analysis four relevant time periods were considered: 10 am, 12 pm, 14 pm and 16 pm. Shade needs in percent were computed with the following equation:

shade needs  =  con[topo shade, 100, height * abs(sin(str orientation - sun azimuth)) * canopy / tan(sun altitude) * stream width]
negative values represent exposed streams, the lowest the values the more need for shade.

        Exposed streams at temperature peak of the day- uncut case

        Exposed streams at temperature peak of the day- cut / burned scenario

For identifying the most impacted streams we identified the values that stay unchanged over the day intersecting the grids of exposed waters at all time periods considered.

        Most shade needs
 

Conclusions

Even though the average DBH is low it can be noticed that a possible harvest on the 50 year or older stands could leave most of the streams exposed.
 

Large Woody Debris Recruitment

Method

This part of the module assesses possible hazards in LWD recruitment for short and long terms.
Determining recruitment hazard is based on the channel sensitivity to wood inputs and riparian recruitment potential.
Channel sensitivity was delineated according to stream gradient and valley confinement using the matrix:
 

grad /confine	< 1%	1-2 %	2-4 %	4-8%
unconf.	low	mod	mod	low
mod. conf.	mod	high	high	mod
confined	high	high	mod	low

table 3. Stream sensitivity to LWD inputs

        Stream sensitivity    1 - low, 2- moderate, 4- high

It can be observed that most of the streams fall in the low category meaning they will not be impacted by wood - consequence of high gradient.
Recruitment potential is strongly influenced by the following: average tree height, stand density, dominant species and presence-absence of wood in streams.
Generally conifers are considered more resistant to decay in streams; a higher stand density means a better chance for recruitment; larger diameter and taller trees are good candidates for potential recruitment. Having no information about presence- absence of LWD both cases were taken into account.
WAM provides classification schemes for the factors above and the matrix for computing the recruitment hazard (tables D-1 to D-5).

        Recruitment potential

Due to low tree height (this is where the data proves inaccurate) the recruitment potential appears as low all across the watershed.
For a better understanding of the watershed we delineated a ratio of stream depth and diameter. The values are very low, trees appear unable to support the fish related processes and could be easily washed downstream.

Results:

        Recruitment hazard LWD on /off - real case

        Recruitment hazard LWD on /off - cut / burned scenario

        Ratio depth / dbh

                                                                                        higher hazards mean lower probability of valuable LWD in streams for the immediate term.

Conclusions

Given the data we had to work with, results showing John's Creek watershed in a poor condition from the LWD point of view are not surprising. A potential harvest of the older stands wouldn't make much difference for wood in streams but would strongly impact shade and stream temperature.
 

References and Acknowledgments

Standard Methodology for Conducting Watershed Analysis (Chapters D and E). WA Forest Practice Board, Version 4.0, November 1997.
Water in Environmental Planning. Dunne Thomas 1943

Put together by Flo Damian  evenflo@u.washington.edu  winter 1999.

Back to top

Back to outline