Effects of green-tree retention on the availability of arthropod prey to bark-gleaning birds: assessment from a large-scale field experiment

Juraj Halaj¹, Maria Mayrhofer², and David A. Manuwal²

¹Cascadien, Inc.
Corvallis, Oregon 97330 USA
jhalaj@cascadien.com

²College of Forest Resources
University of Washington
Seattle, Washington 98195 USA

Green-tree retention is gaining popularity as a forest management tool in the Pacific Northwest. The Demonstration of Ecosystem Management Options (DEMO) study is the first experimental attempt to assess the ecological consequences of this silvicultural model to a wide range of forest organisms. We report on an ongoing study investigating the effects of varying levels (100%, 40% and 15%) and patterns (aggregated vs. dispersed) of green-tree retention on the availability of arthropod prey to bark-gleaning birds. The experiment has a randomized block design and utilizes five 13-ha treatment units at each of three DEMO study sites in the western Cascade Range of Oregon and Washington. Arthropods were collected using crawl traps installed on live Douglas-fir trees and snags and with D-vac sampling of tree boles during bird breeding season from May through August 2003. Early results from June and July 2003 show that the bark arthropod community was dominated by Collembola (38.7%), Arachnida (14.2%), Hemiptera (13.7%) and Coleoptera (11.0%). Arthropod abundance was 1.2-2.3 times higher on live trees than snags but biomass was similar on both categories of trees. Spiders appeared especially sensitive to different combinations of green-tree retention. Significant shifts in the spider guild structure were found among treatments and their abundance on live trees was about 14-32% higher in dispersed than aggregated treatments, and 13-20% higher at 15% than 40% tree retention levels. These findings, which could be explained in part by variation in the forest overstory structure and microclimatic conditions, suggest that the quantity and spatial arrangement of remnant trees after harvest may have considerable impact on the availability and quality of food to bark-gleaning birds.

Introduction

Green-tree retention is gaining popularity as a forest management tool in the Pacific Northwest. This practice is mandated on federally managed lands by the Standards and Guidelines of the Northwest Forest Plan. An explicit assumption of this silvicultural model is the preservation of characteristics of late-successional forests to conserve biodiversity. Effective implementation of this goal, however, requires a thorough understanding of how changes in forest structure affect ecological processes and community development. The Demonstration of Ecosystem Management Options (DEMO) study is the first attempt to provide the needed empirical evidence to evaluate consequences of green-tree retention in the Douglas-fir region of the western United States and provides a unique platform to study species and habitat interactions at scales relevant to forest management.

Figure 1. (A) Remnants of live trees and snags in a harvest-matrix unit (photo C. B. Halpern, University of Washington). (B). Brown creeper, Certhia americana (photo J. Parrish, Utah Division of Natural Resources).

Materials and Methods

The study is a randomized-block experiment utilizing three DEMO blocks (DP, LWS, WF) and five treatments. Treatment units represent 13-ha forest stands (e.g., Fig 1A) with varying levels of retention (100%, 40% and 15%) of the original stand basal area and arrangement of remnant trees (1-ha aggregates vs. dispersed trees).

Figure 2. (A) A bark trap [a modified design after Hanula & New (1996)] and (B) D-vac arthropod collector.

Within each stand, arthropods were collected with crawl traps (Fig. 2A) installed on live Douglas-fir trees (n = 10-20) and snags (n = 10-20) with a dbh of ≥ 50 cm, randomly selected in permanent vegetation plots. Arthropods were sampled at 2 to 4-week intervals during brown-creeper breeding season from May through August 2003 with a total of 540 traps. Less mobile species and stages of arthropods were sampled with a D-vac sampler (Fig. 3B) from the same number of trees in June and July 2003. Collected arthropods were identified to order or family level and their biomass was estimated using length-weight regression models. Data were analyzed with ANOVAs using stand-level averages of response variables. A more detailed analysis of treatment effects on spiders was performed since these predators appear to be an important food resource of brown creepers. A fecal analysis is also being conducted to further clarify dietary preferences of brown creepers in our study system.

Results and Discussion

Figure 4. Guild structure of corticolous arthropods expressed as (A) number of individuals and (B) dry biomass. Based on a total of 37,182 individuals collected in bark traps June-July 2003. Collembola were excluded from subsequent analysis since they are unlikely to represent a significant food source for brown creepers.

Preliminary analysis of bark-trap data shows that the arthropod community was dominated by Collembola (38.7%; 88.3% Entomobryidae), Arachnida (14.2%), Hemiptera (13.7%) and Coleoptera (11.0%) (Fig. 4A). In contrast, the majority of arthropod biomass on tree bark was contributed by Coleoptera (31.9%; 66.3% Curculionidae), Diplopoda (29.3%), Orthoptera (17.8%) and arachnids (15.1%) (Fig. 4B). Spiders comprised as much as 96.9% and 99.8% of arachnid abundance and biomass, respectively.

Figure 5. Effects of retention treatments and tree status (live trees vs. snags) on the abundance of (A,C) total arthropods (excluding Collembola and Arachnida) and (B,D) arachnids collected in bark traps June-July 2003.

Analysis of data from bark-traps showed significant differences among retention treatments, indicating a trend of numerical increase in the abundance of arthropods with the level of tree removal (Fig. 5). With the exception of live trees in 40%A, arthropod numbers increased in all treatments from 12% to as much 158%, compared to control. The highest densities of arthropods were collected in 15%D units (Fig. 5A,C). Spiders appeared especially sensitive among arthropods to different combinations of green-tree retention. Significant shifts in the spider guild structure were found among treatments and their abundance on live trees was about 14-32% higher in dispersed than aggregated treatments, and 13-20% higher at 15% than 40% tree retention levels (Fig. 5B,D). Live trees supported significantly higher densities of non-arachnid arthropods than snags (1.2-2.3x); similar, but less pronounced patterns were observed for arachnids (Fig. 5A,B).

Figure 6. Effects of retention treatments and tree status on the biomass of (A,C) total arthropods (excluding Collembola and Arachnida) and (B,D) arachnids collected in bark traps June-July 2003.

Similar to abundance data, more arthropod biomass was found on trees in the lowest-retention units but this trend was not statistically significant (Fig. 6A,C). In contrast, control stands and 40% retention units had the highest levels of arachnid biomass, which was caused in part by a decline in the abundance of the Amaurobiidae, spider species with the largest body size in our study (≤ 18.8mm), in the highly fragmented 15% treatments. A greater abundance of prey on live trees implies that snags may represent less suitable foraging substrates to brown creepers. Observations of creepers’ foraging behavior in western Oregon also show that the species prefers large live Douglas-fir trees. Interestingly, however, both categories of trees in our study supported similar amounts of arthropod biomass (Fig. 6), suggesting that creepers may show preferences for prey categories found on live trees or that other non-trophic factors influence the selection of their foraging substrate. Results from the ongoing fecal analysis should help us clarify food preferences of brown creepers and focus our analysis on ecologically relevant prey groups.

Figure 7. Examples of the spatial pattern of arthropod distribution on live trees in treatment units (A) 15%D and (B) 40%A of WF block. Study design utilized a 7x9 sampling grid of permanent vegetation plots with a 40-m spacing between plot centers. Each bubble represents values from 1-2 bark traps and its size is proportional to the average amount of arthropod catch (mg) per trap per day, collected June-July 2003.

Spatial patterns of arthropod distribution showed substantial levels of heterogeneity within experimental stands (Fig. 7). This could be related to variation in microclimatic conditions or quality of understory vegetation. In an ongoing analysis, we are using existing vegetation datasets from permanent vegetation plots to test the significance of understory vegetation as predictors of arthropod abundance. In addition, the spatial characteristic of our data also allows the use of a GIS analysis to generate a resource utilization distribution for each treatment unit to asses the relative importance of specific habitat resources to brown creepers. This work is in progress.