sites were sampled weekly during spring high flow (May-June) and twice monthly during 

 summer low flow (July- August), yielding 9 sample dates for most sites. 



Grab samples were collected for nutrients at each site following the protocol 

 described by Ingman (1992a) in order to be consistent with data collected by the MT 

 DEQ. Samples for nutrient analysis were frozen with dry ice in the field and shipped to 

 the Montana State Environmental Laboratory in Helena for nutrient analysis. Analysis 

 included total Kjeldahl nitrogen (TKN), nitrite plus nitrate (N02/N03), total phosphorus 

 (TP), and soluble reactive phosphorus (SRP), which was filtered on site with a .45 um 

 membrane filter. Detection limits for analysis were <0. 1 mg/1 for TKN, <0.01 mg/1 for 

 nitrate/nitrite, and <0.001 mg/1 for SRP and TP. All nutrient sampling equipment was 

 acid washed in 50% instranalyzed HC1 and triple-rinsed in deionized water. Field blanks 

 were prepared for each sampling date for quality assurance. Sample results for TKN and 

 N02/N03 were summed to estimate total nitrogen (TN). Total nutrient loads were 

 estimated using discharge data collected using standard pygmy flow meter. Gardiner 

 ditch (Station 3) and the Dutchman diversion (Station 6) are exceptions since discharge 

 could not be measured and only TN and TP concentrations were determined. 



Temperature and pH determinations were made at each site on each visit using a 

 Orion Model 250A portable pH meter. Turbidity samples were brought to the laboratory 

 and analyzed using a Hach 2100A turbidimeter within 24 hours. Samples were collected 

 for total suspended sediment determination by filtration method. 



A combination of spreadsheet (Microsoft Excel) and statistical software (SPSS) was 

 used to manage and analyze physical and water quality data. Simple descriptive 

 statistics (i.e. means and 95% confidence intervals) were used to generate summary 

 tables and graphs to assess differences between sites. Because initial analysis of water 

 quality data based on flow period (i.e. high spring flow vs. low summer flow) did not 

 reveal any additional significant information, tables and graphs of water quality data are 

 presented in terms of summer (May through August) mean values (See Fig. 1 through 9). 



Riparian inventories were performed using the UM School of Forestry's Riparian and 

 Wetland Research Program's Lotic inventory (detailed inventory). Forms and description 

 of protocols are available online at http://www.nvrp.umt.edu. The study area was 

 divided into areas called polygons, covering approximately 0.5 stream miles and 

 bordered by the edge of the riparian zone. Ending and starting points for polygons were 

 delineated by a combination of GPS coordinates, photo documentation and narrative 

 descriptions. Specific areas of concern (i.e. severely eroding banks, headcuts, etc.) were 

 recorded in a similar manner. Riparian inventories were completed for the entire length 

 of the proposed restoration area (see map), except where the creek entered wetland and 

 beaver complexes above the reservoir. In this area, there was a lack of distinct channel 

 or riparian boundaries so assessments were not feasible. 



Lotic inventories involved recording the presence and coverage of plant species, 

 infestation by invasive species, and age class and utilization of woody species. In 



