Introduction

    Valley Oak (Quercus lobata) is the largest of the world’s oak species.  It is endemic to California, and ranges from the Mount Shasta region of northern California to the Santa Monica Mountains in Southern California.  Decades of published studies have noted a long term decline in Valley Oak regeneration and recruitment throughout the species’s range.  

    The term “recruitment” refers to the transition from seedling to sapling.  The vast majority of Valley Oak seedlings do not survive past the first year, and thus never make the transition from seedling to sapling.  For this reason, researchers are often less interested in instances of Valley Oak germination, and more interested in conditions under which successful recruitment might occur.  In general, when a tree reaches 30 cm in height, it is said to have been “recruited” into the world of being a tree.

    Although many studies have been published which examine the recruitment problem in Northern and Central California, very little research has been published on the subject as it pertains to the species’s Southern-most habitat in the Santa Monica Mountains.  This project focuses on one site (Cheseboro Canyon) in Agoura Hills in the Santa Monica Mountains where Valley Oak recruitment is known to have taken place in recent decades (Figure 1).  The goal of this project is to use GIS Suitability Analysis to determine the existence of any local site-specific factors that could be having an affect on Valley Oak recruitment at the site.  Based on the results obtained, an attempt will be made to predict locations within the site where successful recruitment might be expected to occur in the future.

                 Figure 1.  Cheseboro Canyon study site in the Santa Monica Mountains

Data

    This project examines seventeen Valley Oak saplings that were recruited between 1994 and 2001 (dates were determined by the authors in a previous study).  The specimens’ locations were determined in 2009 by Professor James Hayes of the CSUN Geography Department using a GPS device in the field.

    The aerial photograph of the site was obtained through the Bing Maps Aerial server.  Both the road layer and the streams layers were digitized. 

    A DEM of the Santa Monica Mountains was obtained from the USGS seamless server.  A slope layer was then created using this DEM which depicts slope in degrees.   

Methods

    The initial variables used to perform the site suitability analysis were elevation, slope, and proximity to streams.  

    Upon inspection of the DEM, we noted that all of the specimens were located between 280 and 300 meters above sea level.  We used the reclassify tool to reclassify the DEM, and assigned a value of “1” to locations that fell within the afore-mentioned elevation range; all other locations were assigned a value of “0”.  Similarly, after noting that all of the specimens were located in relatively flat bottom lands with slopes between 0 and 8 degrees, we again used the reclassify tool to reclassify the slope layer (which we created from the DEM using the slope tool), and assigned a value of “1” to all locations with slopes between 0 and 8 degrees; all other locations were assigned a value of “0” (Figure 2).  

                            


                                            a.                                                                              b.
                                   
                               Figure 2. (a) Reclassified DEM layer and (b) reclassified slope layer

    Next, we used the euclidian distance tool to produce a raster layer depicting distance from streams in ten 10-meter classes.  We set the maximum distance to 100 meters.  The results revealed that 15 specimens were within 40 meters of a stream.  Of the remaining 2 specimens, 1 was 60 meters away from any stream, and 1 was 100 meters away.  Based on these observations, we once again used the reclassify tool to reclassify the euclidean distance raster layer, and assigned a value of “2” (most suitable) to cells located between 0 and 40 meters from a stream, and a value of “1” (somewhat-suitable) to cells located between 41 and 100 meters from a stream (Figure 3).

      

          














                                    a.                                                                                                                           b.

Figure 3. (a) Map depicting euclidean diastance away from the streams and (b) reclassified euclidean distance layer.

    Once we had reclassified all the layers, we used the raster calculator to produce a suitability map containing 4 classes of suitability values.  Locations that were most suitable received a value of “4”, and locations least suitable received a value of “1”.  The map revealed that the locations of 15 specimens were within the “4” range; the remaining 2 specimens were located within the “3” range (Figure 4).  

Figure 4.  Site Suitability Map

    After inspecting the site suitability map, it became apparent that large portions of the site which did not contain any young trees received suitability values of “4” (most suitable).  We noticed that the 17 specimens were clustered in a narrow, roughly north-south trending line along one of the streams.  However, we also noticed that young trees were largely absent in areas that received suitability values of “4” surrounding a second stream.  This fact suggested that some undetermined variable other than elevation, slope, or proximity to streams was causing recruitments to occur along one stream but not the other.  

    Based on evidence we uncovered in a previous study, we decided to add “proximity to Cheseboro Road” as a fourth variable to the suitability analysis equation.  Cheseboro Road runs along the north-south trending stream, around which most of the trees were clustered.  We digitized the road based on the aerial photo used as the base layer.  We used the euclidean distance tool to produce a raster layer depicting distance from the road in ten 10-meter classes, with the maximum distance set to 70 meters.  The results revealed that 16 of the specimens were within 20 meters of the road.  The remaining specimen was 70 meters from the road.  Based on these observations, we used the reclassify tool to assign values of “2” to cells located less than 30 meters from the road, and values of “1” to cells located between 31 and 70 meters from the road.

    We again used used the raster calculator to produce a suitability map.  This time we added our fourth variable (proximity to road) into the equation.  The resulting map revealed that cells with the highest suitability values clustered around the road, rather than the streams (Figure 5).  

                    Figure 5. Site Suitability Map with proximity to Chesebro Road included.

Results

    The first suitability map we created (that did NOT contain the road variable) showed that the most suitable land was clustered around the two streams.  The elevations and slopes surrounding the streams all received the highest value of “1”.  However, when we added the “proximity to road” variable to the raster calculator equation and produced a second suitability map, the results were much different.  The land around the north-south trending stream retained its high suitability values, but only because the road followed the course of this stream (Figure 6).  Land surrounding the second stream that had previously been classified as highly suitable was now classified as completely unsuitable.  These results seemed more accurate to us because they corresponded well to the actual locations of the 17 specimens.

   Figure 6. Final site suitability map for Valley Oaks at the Chesebro Canyon study site.

Discussion

    Our analysis suggests that the presence of Cheseboro Road has been an important factor influencing Valley Oak recruitment at Cheseboro Canyon between 1994 and 2001.  16 of the 17 specimens are located within 20 meters of the road, a fact which leads to a discussion of why the road might be having such a large affect on recruitment.

  In a previous study, we looked at aerial photos of Cheseboro Canyon obtained from the CSUN map library and determined that Cheseboro Road was paved in 1994.  When a road is paved, it can no longer absorb water, and rainfall tends to runoff and collect on either side.  This fact, along with road splash from passing vehicles, often leads to increased vegetation along roads, especially in areas where water is a scarce commodity.  Passing cars may also keep animals from coming too close to roads.  If this is the case at Cheseboro, it may be significant, given the fact that numerous studies of Valley Oak throughout its range conclude that predation by rodents and small mammals may limit recruitment at many sites.  

  Many studies often list grazing as another potential limiting factor to Valley Oak recruitment.  Grazing ended at Cheseboro in 1985, and yet the first recruit does not occur until 1994, the year that the road was paved.  This, along with the fact that only 1 recruitment in the data set occurred farther than 20 meters from the road, seems to suggest that Cheseboro Road may be a strong explanatory variable influencing recruitment, and that future recruitments may be expected to occur with similar proximity to the road.  If this is so, it may mean that water that was once available to Valley Oak throughout the site may no longer be available, and thus saplings are increasingly relying on water provided by the affects of road-paving.