Windbreaks and living snow fences are a recognized technology for controlling wind and snow in the Great Plains and Midwest. They are not so common in the dry land wheat-growing region of Eastern Washington. In the past, o­nly about half of the plantings survived in a manner to give good performance. This can attributed to poor planning and inadequate care, leading to the attitude that trees are too difficult to establish. Today, the use of fabric mulch has greatly enhanced windbreak establishment success. Successful living snow fence projects in southeastern Idaho and dry land test tree plantings in Adams County, WA contributed to the development and successful installation of a living snow fence demonstration project in Lincoln County, near Davenport, WA.

Three-year tree growth was measured at this living snow fence research and demonstration project near Davenport, WA. (Photo courtesy G. Kuhn)

The USDA National Agroforestry Center in partnership with the USDA Natural Resource Conservation Service developed a cooperative living snow fence project with key state and local agencies. The purpose of the project was to show the value of living snow fences in this dry cropland region and to demonstrate new establishment technology. Interagency meetings were held the year before the site was selected to inform the local agencies about living snow fence technology and to form an interagency team for project planning and installation. It was stated "Project planning will be more challenging than the actual project planting" (G. Kuhn, personal communication, 2001). Much energy went to identifying snowdrift sites, field reviewing sites, and securing landowner cooperation.

Road maintenance personnel from the Davenport office of the Washington State Department of Transportation identified 16 snowdrift sites o­n state highways radiating out of Davenport, WA. Upon field review, eight sites were eliminated due to insufficient topography and seven sites eliminated due to unresponsive landowners. All of the proposed living snow fence locations had to be placed o­n private cropland to attain proper snow drift control. The project site was selected o­n Highway 25, nine miles north of Davenport, WA. This is a highly traveled road, leading to many recreational opportunities further north. The landowner was supportive of the project and cooperated throughout the planning and installation.

The site was located in dry cropland, usually wheat/barley rotations. The soil was a deep silt loam. Precipitation was 16 inches. Upon soil testing with a soil auger, a chisel pan and was located 18 inches below the surface. Deep ripping with a 36, three prong ripper was done in the future location of the tree rows the fall before planting. A double twin-row high-density design was recommended using Rocky Mountain juniper. The windward twin row was located 150 feet windward (west) of the road. The leeward twin row was located 42, from the windward row. The space between the twin rows was seeded with Durar hard fescue, a low growing bunchgrass for site protection.

The project was installed the first week of April 2002. An interagency crew planted the trees and installed fabric mulch o­n each row. Local and regional news media were o­n hand to cover the installation and interview key personnel o­n the project and living snow fence technology. Each twin row is 900 feet long, with 532 trees planted. High quality 20 cubic inch container Rocky Mountain juniper seedlings were used. This larger stock has been proven superior for evergreen establishment in semi-arid regions of Washington and Idaho. The summer of 2002, in this region, was o­ne of the driest o­n record. Growth measurements taken during the fall of the first growing season indicated that the trees grew average of 20 inches. That was remarkable, and can be attributed to proper planning, site preparation, quality nursery stock, proper planting, and the use of fabric mulch. Due to the initial success of this project, two other living snow fence demonstration projects have been developed for spring of 2005 planting. Both will protect public roadways. o­ne is located near Anatone, WA and the other is located near Athena, OR.

Three Year Growth Responses

Locations of the pairs of twin rows were designated as west (W) and east (E), with east pair of rows being closer to the highway. The individual rows located within each pair of rows were identified by the numerals as row o­ne (1) and row two (2), with row two being closer to the highway, or east of row o­ne. Row E2, then, designates the row that is closest to Highway 25 and row W1 designates the row that is furthest from Highway 25. Each row was further subdivided into three equal segments, North, Middle, and South that were used to assess potential variation in tree development by north-south position within a row.

Tree measurements were taken in October of 2003, 2004, and September 2005. Every fifth tree in each row was measured beginning with the northern most tree and proceeding southward along the row. There were 27 trees sampled in each row, with a total 108 possible measurements for each year: 54 trees for each location, 27 trees for each row within a location, and 9 trees for each position within a row. We had 100% survivorship in both measurement years. For each sampled tree, we measured the total tree height and the crown width, both measured in inches.

The prevailing wind direction at this installation was from the west, making the west location windward and the east location leeward, with row o­ne being the windward row and row two being the leeward row within each twin-row location. In addition to the prevailing wind direction, there was also a slight mound centered along the north-south line within the rows of trees. Given the preferential wind direction and central mound within the rows, we were interested in identifying differences in average tree size between the locations of the paired rows, between the rows within a location, and among the positions within a row for each location. We were also interested in whether there were differences in the average tree size over time, and whether differences would persist as long term trends in average tree size over time within each nesting level.

To identify the differences in average tree size by year and nesting level we performed a nested or hierarchical analysis of variance (ANOVA), assuming fixed effects, using a nested layout with measurement year added as an additional nesting level above location. We considered two characteristics of tree size: total tree height and crown width.

Total Tree Height

Variability Among the Locations (Windward vs. Leeward): The mean height values for the west (windward) location were smaller than the mean height values for the east (leeward) location for the measurement years 2003 and 2004, 22.0-0.68 inches and 39.6-0.91 inches for the west location and 22.8-0.48 inches and 40.6-0.71 inches for the east location, respectively. This relationship was reversed for measurement year 2005, when the mean height for the west location, 54.6-0.97 inches, became greater than the mean height of the east location, 53.5-1.29 inches. No strong or weak potential differences in mean height were identified between the east and west locations for any year.

Variability Among the Rows within a location: Within each location, the windward rows, rows W1 and E1, had smaller mean height values than the leeward rows, rows W2 and E2, respectively, for all measurement years. A strong potential mean height differences between rows within a location was found o­nly in 2004 for rows E1 and E2 of the east location. A weak potential mean height difference was also found in 2003 for the east location, the mean for row E2 was outside the 95% CI computed for row E1.

Crown Width

Variability Within Rows:
Strong potential differences in crown width were identified between the north and middle positions of row W1 and its south position for all measurement years, between the middle position and south position of row W1 in 2003, between the north and south positions of row W2 in 2003, between the middle and south positions of row W2 and its north position in 2004, and in 2005 between the north position and the middle position of row W2 and between the middle position and south position of row W2. Weak potential differences were found between the north and middle positions of row W1 and between the middle and south positions of row E1 and its north position for 2003, and between the north and south positions and north and middle positions of row W2 in 2005.

Variability Among the Locations: Variability among locations was not found to be significant for any measurement years

Conclusions

Over time, the variability in the individual tree measurements increased somewhat for both height and crown width, as indicated by the increasing values of their standard error values. This implies that as the trees get larger, o­n average, the distribution of tree sizes is widening, and some form of size differentiation is taking place, possibly caused by microsite variations.

In summary, we feel that the first three years of the Davenport WA living snow fence demonstration have been a success. We have demonstrated, so far, that agencies and landowners can work cooperatively to establish a successful demonstration planting. In addition, that Rocky Mountain juniper establishes readily o­n these sites and grows rapidly. Initial Rocky Mountain juniper survival and growth was remarkable and we estimate that the living snow fence will reach functionality in approximately 5 years. Initial success has resulted in additional demonstration sites, under different site conditions in the Inland Northwest.

Acknowledgements

The authors would like to recognize the USDA National Agroforestry Center and the USDA Natural Resources Conservation Service for project funding and oversight. We would like to commend the landowner and interagency planning and installation team for their cooperation and diligence in project planning and installation.

References

Hanley D. P. and Kuhn G. 2003. Trees Against the Wind, WSU Extension, PNW0005, 40 pp.

(Condensed from a handout produced for the Living Snowfence Workshop, October 4-5, 2005, Spokane, WA. For a copy of the complete paper, contact Gary Kuhn, This email address is being protected from spambots. You need JavaScript enabled to view it.)

By Gary Kuhn (USDA NRCS), Dennis Robinson (USDA NRCS), Donald Hanley (Washington State University) and Kevin Gehringer (Biometrics Northwest)