The relative importance of factors affecting age-specific seedling survival of two co-occurring oak species in southern California

• Published in: Forest ecology and management 2008-04, Vol.255 (7), p.3063-3074
• Main Authors: Tyler, Claudia M., Davis, Frank W., Mahall, Bruce E.
• Format: Article
• Language: English
• Subjects: age; Browsing; CART; cattle; classification; Classification and regression tree; herbivores; Herbivory; Inter-annual variation; Limiting factors; Mediterranean; mortality; natural regeneration; Oak recruitment; Quercus agrifolia; Quercus lobata; regression analysis; Seed predation; seedling growth; seedlings; small mammals; temporal variation; trees
• ISSN: 0378-1127; 1872-7042
• DOI: 10.1016/j.foreco.2008.01.073

As has been reported in other oak habitats throughout the Northern Hemisphere, natural recruitment rates of young oaks in California are very low for some species and in some regions. The majority of experimental studies that contribute to our understanding of this oak recruitment pattern in California have been relatively short-term, conducted on a small-scale. Thus, while we have valuable information about the array of factors that are able to limit seedling establishment, we know much less about their relative importance, and how they vary across years, sites, or age classes. To investigate the impacts of factors limiting seedling and sapling establishment of valley oak ( Quercus lobata) and coast live oak ( Quercus agrifolia) across a landscape and over time we replicated experiments in four different years in Santa Barbara County, California. Experimental factors manipulated included cattle grazing, and access by deer and small mammals. Plots were distributed across three sites with a cumulative area of 200-ha, and seedling survival was recorded for a minimum of 5 years. We used classification and regression tree analysis (CART) to examine the relative importance of factors influencing survival of planted oaks at different life stages. The relative importance of limiting factors varied among age classes for both species. For initial seedling emergence and survival to 6 months planting year was the most important factor and rodent access was the second most important factor for both Q. lobata and Q. agrifolia. For survival of seedlings through their 1st year rodents, planting year, and site were major limiting factors, though their relative importance varied between the two species. For survival from 1.5 to 5 years, Q. lobata was strongly affected by rodents and site, while Q. agrifolia was mainly affected by site, deer browsing (which reduced survival), and indirect effects of winter–spring cattle grazing (which improved survival). Contrary to our expectations, based on observed patterns of natural recruitment, Q. agrifolia had equivalent or lower survivorship than Q. lobata in all seedling age classes. Overall, in addition to other factors, there were strong year and site effects controlling oak seedling emergence and survival and the relative importance of limiting factors depended on seedling age. Our results suggest that large-scale, long-term experiments encompassing multiple sites will improve our understanding of controls on recruitment in oak