Determinants of Acorn Productivity Among Five Species of Oaks in Central Coastal California

We measured acorn production of five species of oaks (genus Quercus) over a ten year period (1980-1989) at Hastings Reservation in Monterey County, California. Crop production was highly variable and generally asynchronous between species. Variance in crop size decreased directly with increasing species diversity across sub-areas within the study site. Sources of variance in mean annual crop size include annual and interspecific differences and extrinsic factors correlated with different sub-areas of the study site. However, the majority of variance remains unexplained after controlling for these factors and is apparently due to individual variation in crop production. Climatic variables were not found to correlate with mean annual crop productivity in any of the species, with the possible exceptions of Q. chrysolepis and Q. kelloggii. However, strong autocorrelations were found within the crop sizes of individuals indicating the existence of masting cycles on the order of three years for Q. douglasii and Q. agrifolia, two to three years for Q. lobata, and possibly six years for Q. chrysolepis and Q. kelloggii. These results indicate distinct reproductive strate­ gies in California oaks; continued studies will be necessary in order to identify the relative importance of the evolutionary factors that have resulted in these patterns. Acorn crop size is of fundamental importance both to the understanding of oak regeneration problems and to the diverse assemblage of wildlife dependent on acorns as food. However, surprisingly little is known concerning the factors determining acorn crop size or variability. This ignorance is acutely evident in California, where oaks dominate millions of hectares of land area (Griffin and Critchfield 1972, Griffin 1977, Bolsinger 1987), and acorns are a primary food resource for a wide variety of wildlife (Barrett 1980, Verner 1980). The difficulties of assessing the factors important to crop size are formidable. Acorn crops are notoriously variable, not only at the species level but also at the community, site, and individual tree levels. Moreover, the pattern of spatial heteroge­ neity in crop sizes is likely to be complex given the diverse geographic and climatic range inhabited by oaks in California. Finally, the logistic problems associated with reliably assessing acorn crop size are considerable. Short-term data collected by Presented at the Symposium on Oak Woodlands and Hardwood Rangeland Management, October 31-November 2, 1990, Davis, California. Associate Research Zoologist, Post-doctoral Research Associate, and Graduate Student, respectively, Hastings Reservation, University of California, Carmel Valley, California; and Assistant Professor, Department of Biology, Allegheny College, Meadville, Pennsylvania. Ronald L. Mumme different observers are unlikely to be of value unless carefully coordinated. Acorn traps can provide an accurate estimate of acorn fall for individual trees, but unavoidably miss acorns removed prior to maturity by birds and other animals, require considerable maintenance, and are difficult to put out in suffi­ cient numbers to yield satisfactory sample sizes. Visual esti­ mates (Graves 1980, McKibben and Graves 1987) are subject to considerable variation depending on the observers but have the potential advantage of yielding large sample sizes with reasonable time investment. With these difficulties and challenges in mind, we began a study in 1980 of the fruiting patterns of five species of sympatric oaks at Hastings Reservation in Monterey County, California. Our goal in this continuing endeavor is to quantify the patterns of intraspecific, interspecific, annual, and spatial variability in acorn production and to determine the environmental and intrinsic factors responsible for the observed variation. Analyses based on the first seven years of data were presented earlier (Carmen et al. 1987). Here we extend our earlier results using data through 1989 (10 years), emphasizing the masting patterns observed at Hastings and the factors, both extrinsic and intrinsic, contributing to individual variation in acorn productivity. We also discuss some of our results concerning local variation in mast produc­ tion among and within species and the implications of our findings to the understanding of the relationships between acorn producers and their predators. STUDY AREA AND METHODS Hastings Reservation is a 900 ha reserve located in the Santa Lucia Mountains of central coast range of California, approxi­ mately 42 km inland. Elevation ranges from 467 to 953 m. Mean annual temperature is 17°C. Annual rainfall ranges from 26.1 to 111.2 cm, with a 50-yr mean of about 55 cm. Topography includes relatively narrow valleys amid rugged foothills. Com­ mon at the lower elevations are Quercus lobata, Q. douglasii, and Q. agrifolia, while at higher elevations Q. kelloggii and Q. chrysolepis are also present. These five species include oaks of all three subgenera (Quercus: Q. lobata, Q. douglasii; Erythrobalanus: Q. agrifolia, Q. kelloggii; and Protobalanus: Q. chrysolepis), both evergreen (Q. agrifolia and Q. chrysolepis) and deciduous (Q. lobata, Q. douglasii, and Q. kelloggii) spe­ cies, and species that require one year (Q. lobata, Q. douglasii, and Q. agrifolia) and two years (Q. kelloggii and Q. chrysolepis) for acorns to mature. 136 USDA Forest Service Gen. Tech. Rep. PSW-126. 1991 Beginning in 1980, we established transects in foothill woodland, savanna-grassland, and riparian woodland vegeta­ tion types (Griffin 1983). Along these transects we sampled 250 oaks including 87 Q. lobata, 63 Q. agrifolia, 57 Q. douglasii, 21 Q. kelloggii, 21 Q. chrysolepis, and 1 obvious Q. lobata x Q. douglasii hybrid. Sample sizes decreased slightly in later years as several trees died and were not replaced. DBH of each tree was measured. Each autumn at the height of the acorn crop and prior to acorn fall, generally between mid-September and early Oc­ tober, we measured the relative abundance of acorns of each of the 250 individual trees using two methods. First, two observers scanned different areas of the tree's canopy and counted as many acorns as possible in 15 s. These counts were added to yield "acorns per 30 s". Second, the two observers agreed on a score between 0 and 4 describing the overall size of the acorn crop for each tree. This score, modified from the visual estimate scale pioneered by Graves (1980), was as follows: 0 (no acorns), 1 (a few seen after close scrutiny), 2 (a fair number, acorns seen readily), 3 (a good crop), and 4 (a bumper crop, acorns just about everywhere on the tree). At least one of the original observers has participated in the annual counts each year since we initiated the study. Based on elevation and habitat, we divided the 250 trees into four sub-areas: (1) the Arnold, a high-elevation area located in the southern end of the reservation (all five species present, elevation 780 to 878 m), (2) the Road, a low-elevation area traversing the center of the reservation (all but Q. kelloggii present in good numbers, a semi-permanent creek running through the area, elevation 470 to 512 m), (3) Big Creek, a lowelevation area running northeast of the Road with a seasonal creek running through it (only Q. lobata, Q. douglasii, and Q. agrifolia present, elevation 488 to 579 m), and Haystack/School Hill, two higher-elevation hills located on either side of Big Creek (only Q. lobata, Q. douglasii, and Q. agrifolia present in good numbers, elevation 579 to 637 m). All four areas are less than a km in diameter and are within 4 km of each other. Weather data was taken from a station permanently located at the headquarters near the northern edge of the reservation in the "Big Creek" sub-area. Included in the weather station during the study period was a rainfall gauge, a max-min thermometer, and a Campbell Stokes-type sunshine recorder. STATISTICAL ANALYSES Using the first year of data (1980), the Spearman rank correlation between the two different measures of acorn pro­ ductivity is 0.963 (n=250, P<0.001); thus, the choice between them is largely a matter of taste. In this paper we have used the number of acorns counted in 30 s (henceforth "N30"), which is a continuous, interval measure, in most of our analyses. In order to reduce the dependence of the variance on the mean, N30 was logarithmically transformed (ln(N30+1)) prior to parametric procedures. The 0 to 4 categorical score was used in the weather and the autocorrelation analyses because we suspect it is less sensitive to tree size, foliage density, acorn color, and acorn size, and thus is probably more comparable across years than N30. RESULTS AND DISCUSSION Pattern and Sources of Variation Acorn production varies considerably both from year to year and from species to species (figure 1). Indeed, there are no significant correlations between the mean yearly crop scores of any of the species except for between the closely related Q. lobata and Q. douglasii (table 1). There is, however, a tendency for the species requiring a single year to mature acorns to be out of synchrony with those requiring two years: all six correlations of mean crop scores between the one-year species and the two-year species are negative, while all four correlations between species within these categories are positive (Fisher exact test, P < 0.05; table 1). One important consequence of the asynchrony of acorn production across species is that variability in masting patterns are dampened. This is shown by comparing the annual variation of the individual species (figure 1) with the mean acorn crop score of all individuals combined (figure 2). Annual variation measured by the coefficient of variation for the mean annual score of all species combined was considerably less (34 percent) than that for any of the five individual species (range 55 to 128 percent). This decrease appears to be a direct consequence, at least in part, of increased species number, as indicated by the relationship between the coefficient of variation in the mean annual crop score of all trees within sub-areas of the study site and the number of species of oaks commonly present in those sub-areas (figure 3a). Values decrease from a mean of 63 percent in the sub-areas with three species to 36 percent in the sub-area with five species. For acorn-dependent wildlife, this asynchrony acts to moderate the otherwise potentially disastrous effects of poor acorn years. These data indicate that annual, local, and interspecific variation in acorn production all contribute importantly to com­ munity-wide patterns of crop size. However, all these factors are apparently overshadowed by individual variation. This is shown by a three-way analysis of covariance (ANCOVA) in which year, sub-area, and species were analyzed as main factors and diameter at breast height (DBH) included as a covariate (table 2). Each factor and the covariate are analyzed controlling for all other variables. All variables and their interaction terms are highly significant. However, in total, they account for less than half (46.1 percent) of the variance in crop scores present in the data. Of the individual variables, most important appears to be yearly variation, while sub-area, species, and DBH of individual trees account for only a small proportion of the total variance. USDA Forest Service Gen. Tech. Rep. PSW-126. 1991 137 Figure 1—Mean (±S.E.) number of acorns counted per 30 s (log transformed) for each of the five oak species at Hastings for the ten years 1980-89. Table 1—Spearman rank correlations of acorn production by five species oak oaks using mean annual data (mean number of acorns counted per 30 s; see text). N =10 years. * = P < 0.05. Q. lobata Q. douglasii Q. agrifolia Q. chrysolepis Q. douglasii 0.83* — — — Q. agrifolia 0.14 0.43 — — Q. chrysolepis -0.42 -0.46 -0.18 — Q. kelloggii -0.18 -0.41 -0.39 0.21 Figure 2—Mean number of acorns counted per 30 s (log transformed) for all five species combined. For each year, values were averaged for the five species to yield a community-wide estimate of overall acorn produc­ tion. Figure 3—Coefficient of variation in the mean number of acorns counted per 30 s (log transformed) for the four sub-areas of Hastings Reservation, graphed according to the number of species commonly present in those sub-areas. Two sub-areas have three common species while one each have four and five. For the sub-area with three common species, the mean ± S.E. is plotted. However, the significant interaction terms make it impossible to determine with confidence the importance of the individual factors, since their effects are not independent of each other. This analysis indicates that individual variation is responsible for a larger proportion of the total variance in crop size than variance associated with different years, species, and sub-areas com­ bined. Only a small part of this individual variation appears to be correlated with tree size. 138 USDA Forest Service Gen. Tech. Rep. PSW-126. 1991 Table 2—Three-way analysis of covariance of number of acorns counted per 30 s (log transformed) using species, year, and sub-area within the study site as main factors and DBH as a covariate. Each variable is considered controlling for all other variables. *** = P < 0.001. Percent of total Sum of squares variance df F -value Main factors Year 436.8 — 9 29.3*** Area 120.1 — 3 24.2*** Species 51.7 — 4 7.8*** Interaction Species x year 1388.1 — 36 23.3*** Area x year 169.1 — 27 3.8*** Species x area 121.3 — 9 8.1*** Covariate DBH 22.5 — 1 14.2*** Explained 3374.1 46.1 89 22.9*** Residual 3950.8 53.9 2383 — Total 7325.0 100.