David R. Peart

Dealing with death data: individual hazards, mortality and bias

In ecology and evolution, we have barely begun to tap the information available in survival data. Who lives or dies, and why, is a large part of natural selection. In ecology, these are key questions for building better individual-based models of population and community dynamics. Powerful analytical tools exist to answer them, but the literature is scattered across disciplines, and its relevance is often obscured by inconsistent terminology and technical presentation. Here, we evaluate methods for the application of such tools to ecology and evolution. Analyses based on individual hazards of death are particularly promising, especially in combination with improvements in sampling design. The same methods can also reduce the largely unrecognized biases that plague population-level estimates of mortality rates.

Effects of hurricane damage on individual growth and stand structure in a hardwood forest in New Hampshire, USA

1. Growth responses to hurricane disturbance were assessed for the dominant species in two sites in a northern-hardwood forest at Hubbard Brook, New Hampshire, USA. The species comprised two shade-tolerant trees, Acer saccharum (sugar maple) and Fagus grandifolia (American beech), one relatively shade- intolerant tree, Fraxinus americana (white ash) and one tree of intermediate tolerance, Betula alleghaniensis (yellow birch). The two sites, 0.6 km apart, were similar except for contrasting disturbance histories. The hurricane-damaged site was devastated by a hurricane in 1938, whilst the control site was only lightly damaged by the hurricane

DENSITY-DEPENDENT POPULATION DYNAMICS OF A DOMINANT RAIN FOREST CANOPY TREE

Intraspecific, negative density dependence may contribute to the maintenance of diversity by limiting the dominance of common species. Shorea quadrinervis Slooten (Dipterocarpaceae) is one of the dominant canopy trees in a species-rich tropical rain forest in Southeast Asia. We test whether juvenile density and performance and overall population growth rate of S. quadrinervis decline with increasing local abundance of conspecific adult trees. We mapped the 357 S. quadrinervis adults (≥15 cm dbh) in 75 ha at Gunung Palung National Park, Indonesian Borneo, and compared the dynamics of juveniles <1 cm dbh between eight areas (80 m diameter plots) with high densities of conspecific adults and eight containing low densities. Within these 16 areas, we mapped 5215 S. quadrinervis juveniles and followed their growth and survival for two years. Plots with high and low density of S. quadrinervis adults did not differ in understory light, slope, aspect, or soil properties nor in the total density (all species combined) of either adults or juveniles. We evaluated four kinds of evidence. All were consistent with the hypothesis of negatively density-dependent population dynamics: (1) Density and size: Although the density of juveniles 75 cm tall. This suggests positive density-dependent seedling recruitment followed by negative density-dependent juvenile dynamics. (2) Foliar condition: Juveniles had more leaves per unit height where adult density was low. Growth and survival increased more with leaf number than with height. (3) Juvenile dynamics: Growth and survival were higher where adult density was low. Previous work suggests that herbivores may drive this negative density dependence in juvenile performance. (4) Population dynamics: We modeled population growth rate (λ) for each of the 16 areas using separate matrix models parameterized with empirical juvenile transition rates. We found that λ declined with local adult density (r2 = 0.40), and this trend was robust to the inclusion of El Nino conditions. We conclude that replacement of adults in the canopy is less likely in areas where conspecific adults are more abundant, tending to limit the dominance of S. quadrinervis in the canopy.

Distance-dependence in herbivory and foliar condition for juvenile Shorea trees in Bornean dipterocarp rain forest

Abstract We assessed density- and distance-dependence in herbivore effects and juvenile condition for four species of Shorea, the most speciose genus in the dominant canopy family of southeast Asian rain forest trees (Dipterocarpaceae). Herbivore damage was quantified as partial leaf loss on young leaves, and whole plant foliar condition as the product of the fraction of leaf nodes containing leaves and the fraction of tissue remaining on extant leaves. Adults of the four species were centers of high total, as well as conspecific, density of juveniles (<1 m tall). For two species, S. hopeifolia and S. pinanga, herbivore damage declined significantly with distance, decreasing by 40% and 51% respectively, between 5 m and 35 m from the parent. For the same two species, foliar condition improved significantly between 5 m and 35 m, increasing by 45% for S. hopeifolia and 24% for S. pinanga. If foliar condition influences juvenile survival and growth, more widely dispersed seeds of these species are more likely to recruit to the canopy. In contrast, there was no significant distance-dependence for S. parvifolia or S. longisperma. Among species, herbivore damage was greatest in those species with greatest local juvenile abundances, i.e., those with highest densities, leaf size, juvenile foliar mass and/ or foliar mass/m2 ground area, but was unrelated to the toughness of mature leaves. However, distance was a better predictor of herbivore damage than was conspecific juvenile density, as evaluated by backward elimination regressions, for both S. hopeifolia and S. pinanga. For foliar condition, the best predictor was distance for S. pinanga, but conspecific density for S. hopeifolia, whose juveniles were smallest and occurred at the highest densities. Total juvenile density (all woody plants) was eliminated as a factor in all cases. The species-specificity of effects (i.e., their dependence on conspecific distance or density), together with the marked differences among congeneric species, caution against generalizations regarding distance-dependent effects in diverse forests.

Understory herb assemblages 25 and 60 years after clearcutting of a northern hardwood forest, USA

Abstract The abundance of understory herbs in 25- and 60-year old clearcuts was compared to adjacent old secondary stands to infer how herb assemblages change in space and time following clearcutting of secondary forest. Herb densities and environmental conditions were sampled along 100 m transects crossing the clearcut/old secondary forest boundary, at three sites with 25-year old clearcuts and three sites with 60-year old clearcuts. At the 25-year sites, species densities of the 23 most frequent species were compared between clearcuts and adjacent old secondary forest to classify species into response groups. Species were classed as “insensitive” if there was little or no difference in density (seven species), “sensitive” if densities were lower in the clearcuts (six species), “enhanced” if densities were higher in clearcuts (four species), and “edge-enhanced” if densities were highest near clearcut edges (six species). Densities of two of the six sensitive species declined significantly with distance from the edge into the clearcut. Further, when regression results for all 23 species were combined, the mean slope of density vs. distance was significantly negative, indicating an overall trend to lower density with distance into the clearcuts. Most species classed as sensitive at 25 years did not show similar sensitivity at the 60-year sites; only one species had lower density in 60-year clearcuts than adjacent old secondary stands. Overall, there were substantial residual effects of clearcutting on herb assemblages within 50 m of the historical edge at 25-year sites, but not at 60-year sites. This difference may be due to different logging practices at the two historical times, as well as the difference in time for recovery since logging.

High abscission rates of damaged expanding leaves: field evidence from seedlings of a Bornean rain forest tree

Herbivore damage is known to cause the premature loss of mature leaves. However, the effects of herbivory on abscission during the early stages of leaf development remain unexplored, even though herbivores frequently prefer unlignified, immature leaves. In a field experiment, we removed 50% of the tissue from leaves at various stages of development on seedlings of Shorea hopeifolia (Dipterocarpaceae), a dominant rain forest tree in Indonesian Borneo. Four weeks following simulated herbivory, >88% of unlignified expanding leaves had been abscised, compared to only 20% of fully expanded, unlignified leaves and 0% of recently lignified, mature leaves. In a separate experiment over 9 wk, simulated herbivory did not increase abscission rates of mature leaves, even when 75% of leaf tissue was removed. Because most (58%) of S. hopeifolia seedlings under natural conditions had lost 1% or less of the tissue from their mature leaves, herbivore damage probably has little effect on the abscission of mature leaves. In contrast, the tendency for damaged expanding leaves to abscise may explain why 49% of S. hopeifolia seedlings had already lost their youngest leaf. If similar patterns occur in other species, herbivore attacks on developing leaves may contribute substantially to both leaf loss and the cumulative impact of herbivory on the growth and survival of whole plants.

Effects of logging history and hurricane damage on canopy structure in a northern hardwoods forest

PEART, D. R., C. V. COGBILL AND P. A. PALMIOTTo (Department of Biological Sciences, Dartmouth College, Hanover, NH 03755). Effects of logging history and hurricane damage on canopy structure in a northern hardwoods forest. Bull. Torrey Bot. Club 119: 29-38. 1992.Canopy structure following the 1938 hurricane and subsequent salvage logging was evaluated from 1942 aerial photographs for six gauged watersheds in the northern hardwoods forest at the Hubbard Brook Experimental Forrest, N.H. Three of the six watersheds that have not been cut since the hurricane were also evaluated from 1978 aerial photographs. A total of 1635 points were sampled on a grid system and categorized as high canopy, low canopy, exposed understory or open, as assessed through a stereoscope. Archival reports on logging and hurricane effects were used to supplement the photographic interpretations. In 1942, percent openness (area whithout canopy cover) ranged from 13% to 38%. From 1942 to 1978, percent openness decreased by a mean of 18%. The reduction in canopy cover due to the hurricane and salvage logging was apparently less than 20% in the gauged watersheds. However, damage was patchy: the most heavily damaged area (72% open) was that which was logged least before the hurricane, and watersheds with most evidence of young second growth prior to the hurricane were apparently the least damaged. Although effects were locally severe, large areas were apparently only lightly damaged in 1938. Recovery of the canopy in the heavily damaged area was rapid: in 1978, only 9% of the area was open, and canopy structure was similar to that in the gauged watersheds. Thus, 40 years of canopy development effectively masked the patchy effects of hurricane damage. Photographic evidence, together with other historical records, provides a better assessment of past disturbance to forest canopies than unquantified eyewitness accounts, and can be evaluated more rapidly and over larger areas than more intensive, field-based investigations.

Effects of competition, herbivory and substrate disturbance on growth and size structure in pin cherry (Prunus pensylvanica L.) seedlings

We examined separate and interactive effects of intraspecific competition, vertebrate browsing and substrate disturbance on the growth and size structure of pin cherry (Prunus pensylvanica L.) in the first two seasons of growth after clearcutting, in a hardwoods forest in New Hampshire, United States. Over the 15-month study period, 97.5% of 1801 individuals survived, and mean plant height increased from 4-fold at high density to 5-fold at low density. Relative height growth was significantly lower at higher plant densities in two of the three growth periods examined. Vertebrate browsers (moose and deer) significantly preferred taller plants. Browsed plants had higher relative height growth following browsing than unbrowsed plants (compensatory growth) at low and intermediate densities. The degree of compensation declined with density and compensation was not significant at the highest density level. At low and intermediate densities, plants browsed early in life regained height dominance through compensatory growth; they failed to regain dominance at high density. Because compensatory growth tended to offset the effects of size-selective browsing, there was no difference in the degree of size inequality between browsed and unbrowsed plots. However, size inequality increased with plant density. Substrate disturbance caused by logging had no significant effects on either relative height growth or size inequality. The slope of the relationship between relative height growth and initial height increased significantly with density and time, and was higher in unbrowsed than in browsed plots, suggesting that competition among plants was size-asymmetric. Despite the preference of browsers for large plants, there was a clear net growth advantage for plants of large initial size, when the effects of competition, browsing and compensatory growth were combined. The interactive effects of density and browsing demonstrate the importance of a multifactorial approach to the analysis of individual plant performance and population structure.

The radial growth response of American beech (Fagus grandifolia) to small canopy gaps in a northern hardwood forest

POAGE, N. J. (Department of Forest Science, Oregon State University, Corvallis, OR 97331 5705) AND D. R. PEART (Department of Biological Sciences, Dartmouth College, Hanover, NH 03755). The radial growth response of American beech (Fagus grandifolia) to small canopy gaps in a northern hardwood forest. Bull. Torrey Bot. Club 120: 45-48. 1993.-American beech (Fagus grandifolia, Ehrh.) is the most shade tolerant of the canopy dominant tree species in the northern hardwood forest at Hubbard Brook, New Hampshire. Nevertheless, subcanopy beech trees responded significantly to small canopy gaps (28-103 m2) by increasing radial growth compared to shaded trees. Gap-edge canopy trees did not respond to canopy gaps. Radial growth was significantly and positively correlated with DBH for shaded subcanopy trees; this relationship was much weaker for subcanopy trees in gaps. That even the canopy dominant species with the most extreme shade tolerant strategy responds substantially to canopy gaps reinforces the view of the central role of gap formation in forest dynamics.

Low mortality in tall tropical trees

The dynamics of the tallest trees in tropical forests are of special interest due to their carbon content, canopy dominance, and the large canopy gaps created when they die. Known ecological mechanisms that may influence tall tree survival lead to conflicting predictions. Hydraulic stress and exposure to high winds and desiccation should increase death rates, yet the tallest trees have the greatest access to light and escape damage caused by falling boles and branches. The uncertainty in tall tree mortality rates has been difficult to address due to their low density, which makes mortality rates challenging to estimate accurately. Here, we use a combination of LiDAR remote sensing and field measurements to show that the mortality rate over 8.5 years among individuals .40 m tall in 444 ha of lowland Neotropical rain forest was 1.2% per year, less than half the landscape-scale average for all canopy trees (2.7% per year). The low mortality is likely explained by species-specific traits that decrease the mortality risk and/or ecological advantages of height that outweigh the risks. Regardless of the mechanisms, the low mortality rate has important implications for tropical forest carbon budgets, as we estimated that a single tall individual represents 2-11% of total live aboveground carbon stocks per hectare. Our findings suggest that height-specific dynamics may be surprisingly different from traditional diameter-specific dynamics, emphasizing the importance of extending ecological studies to investigate the role of tree height in forest dynamics.