Finding Goldilocks: a survey of the factors limiting natural oak recruitment.

This project is innovative because, whereas past studies have crossed and combined a multitude of factors assumed to limit recruitment in order to propose better land management strategies, the practicality of the suggested practices is often incompatible with large-scale impacts because of costly interventions (irrigation, tubes, fencing). These interventions are essentially creating micro-sites favorable to oaks, but there is a gap in the understanding of why micro-sites already exist on many cattle ranches where oaks recruit naturally. Furthermore, because blue oaks are so slow growing, initiating a study to control for various variables doesn't prove that seedlings can make it into the critical sapling stage and onto maturity. Our study looks at the natural experiment which has already taken place and shown us where oaks recruit naturally. Therefore, our only job is to tease out the factors that make one microsite preferable and another microsite next door not.

Because our study took place along a continuum of grazing pressure, we were able to assess site qualities that promote natural oak regeneration. From these results we developed suggestions for grazing management practices. Ultimately, the goal is to help landowners take advantage of the specific conditions on their ranches that encourage recruitment without expensive intervention.

For this study, we secured funding for a baseline of 15 ranches. Adding this SARE grant allowed us to reach 9 more ranches with identified recruitment microsites. This provided us with critical redundancy in our data. Because so many factors are at play, the more microsite replicates we had, the more we were able to tease apart the various factors limiting recruitment. Indeed, even these data, over 60 control and 60 recruitment sites, was at the low end of what was necessary.

For our purposes, we defined a ‘recruitment area’ as any site containing individuals of sapling size (>140cm) with a density greater than 5 individuals per 20 square-meters. A sapling was any individual with a basal diameter of one centimeter or more and a height >140 cm and <6m, based on research suggesting that these thresholds are common with survival through the sapling stage (Sweiki and Bernhardt 1998,  Phillips et al 2007). We used an already awarded NRCS Conservation Innovation Grant to survey ranches for suitable microsites. With microsites identified, we returned to collect data from each site. SARE funding allowed us to expand this experiment to additional ranches and additional microsites providing more strength to identified patterns as well as adding a workshop.

Methods and Objectives:

1. Identify landowners.
   • Working with our partners at local RCDs and NRCS offices, along with several non-profit partners (Audubon California, Point Blue Conservation Science), as well as rancher referrals, we developed a list of willing landowners. We estimated at least two recruitment sites per ranch based on previous observations. In order to build statistically significant models, we located and surveyed 60 individual recruitment locations. Reaching a large number of ranches (defined by property ownership), allowed for large geographic differences and ensured at least 60 sample locations.
2. Set up survey
   • Within the participating ranches we relied on the landowners’ knowledge, personal experience, satellite imagery, and field scouting to identify ‘recruitment areas’ that fit our definition for natural recruitment.
   • In recruitment areas, we measured many factors (vegetation, rockiness, slope, aspect, shade, number of mature trees, count of dung) including a destructive sample to age the small trees. We noted if the recruitment area had multiple age classes (in particular the rare sapling and young tree stages), in this way we know that not only do oaks recruit in those areas, they thrive into maturity.
   • Following the methods of Sweiki et al. 1993, we used the presence of particular size classes as the key determinant for a ‘recruitment area.’
   • Recruitment areas varied in size. To ensure that sampling didn't over report the same phenomena, replicates were at least .25 miles apart, to minimize spatial auto-correlation.
   • Each ‘recruitment replicate’ was paired with a ‘non-recruitment replicate.’ Non-recruitment replicates were located at the nearest site where no recruitment was visible. We repeated all measurements at both recruitment and non-recruitment sites.
3. Measurements
   • Landowner interview: land use history: grazing, discing, other events such as fires.
   • Document water and supplement (e.g. salt lick) locations
   • Create a Cumulative grazing score (c.f., Swieki et al 1997): sum of
     (months grazed) × (relative stocking) × (season factor) for the period 1990-2022. Relative stocking ranges from 0 (none) to 3
     (high). Season factor is 1 for winter and 2 for summer or year-round.
   • Dung counts, as a measure for grazing pressure, taken by transect survey after the rainy season (c.f., Young et al 2005). This will help determine if static barriers (fences, rocky outcroppings) are impacting cattle pressure.
   • Percent bare ground, rock cover, and plant cover by visual estimate.
   • Shade using a percentage (full sun, partial sun, partial shade, full shade).
   • Presence/absence and species of potential nurse plants.
   • Proximity to mature oak tree using a range finder
   • Soil type via NRCS soil map (https://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm).
   • Slope via percentage estimate
   • Aspect by compass app on smart phone
   • Confounding factors: proximity to road or ditch, salt licks or water troughs, presence of browsing obstacles (relic fences, debris, rock outcroppings), or other anomalies.
4. Analysis
   • Statistical analysis will take place in collaboration with the UC Davis Graduate Group in Ecology, and will include ... [multivariate correlation techniques, and model building with Akaike's information selection criterion (AIC)]