Data from: Antagonistic biotic interactions mitigate the positive effects of warming on wood decomposition
Data files
Dec 05, 2024 version files 3.39 KB
-
README.md
784 B
-
Warren.etal.data.csv
2.60 KB
Abstract
Global change drivers such as habitat fragmentation, species invasion and climate warming can act synergistically upon native systems; however, global change drivers can be neutralized if they induce antagonistic interactions in ecological communities. Deadwood comprises a considerable portion of forest carbon, and it functions as refuge, nesting habitat and nutrient source for plant, animal and microbial communities. We predicted that thermophilic termites would increase wood decomposition with experimental warming and in forest edge habitat. Alternately, given that predatory ants also are thermophilic, they might limit termite-mediated decomposition regardless of warming. In addition, we predicted that a non-native, putative termite-specialist ant species would decrease termite activity, and consequently decomposition, when replacing native ants. We tested these hypotheses using experimental warming plots (~2.5°C above ambient) where termites, and their ant predators, have full access and vary in abundance at microscales. We found that termite activity was the strongest control on decomposition of field wood assays, with mass loss increasing 20% with each doubling of termite activity. However, both native and non-native ant abundance increased with experimental warming and, in turn, appeared to equally limit termite activity and, consequently, reduced wood decomposition rates. As a result, experimental warming had little net effect on the decomposition rates – likely because, although termite activity increased somewhat in warmed plots, ant abundances increased more than five times as much. Our results suggest that, in temperate southern U.S. forests, the negative top-down effects of predatory ants on termites outweighed the potential positive influences of warming on termite-driven wood decomposition rates.
README: Antagonistic biotic interactions mitigate the positive effects of warming on wood decomposition
Data file description:
- The data file included in the submission is Warren.etal.data.csv
Variables:
Ident = line number
Treatment = experimental warming treatment (Warming = warming treatment; ambient = control)
Habitat = habitat type (Gap, Forest) Block = experimental block Year = year of data collection
Nest.g = Starting nest weight in grams
Nest.dry.wgt = final nest weight
Nest.perc.soil = percent mineral soil in final nest weight
Ant.sp = ant species (BRCH = Brachyponera chinensis; APRU = Aphaenogaster rudis)
Ant.pres = ant presence/absence
Ant.abund = ant abundance Term.pres = termite presence.absence
Term.abund = termite abundance
Missing data: NA
Methods
Twenty-four study plots were established, using a random blocked design, in a 2,500-m2 area at Whitehall Forest in 2009, with 12 plots located within forested habitat (hereafter “forest plots”) and 12 located in an adjacent manually cleared area at the edge of a powerline cut (hereafter, “edge plots”). The minimum and maximum distances between plots was 2 and 60 m. We used both edge and interior plots because temperate forests have, for a long time, been impacted strongly by human development, resulting in forest landscapes composed of a great deal of edge as well as interior forest which vary in their temperature regimes. Further, edge environments may facilitate non-native ant invasions (Ness 2004; King and Tschinkel 2016; Warren II et al. 2020).
In 2009, resistance heating cables were buried 10-cm deep and 20-cm apart in 18 of the 24 plots, which were built into 18-m2 rectangular (4 x 4.5 m) open-top chambers (1.5 m height) sided with greenhouse PVC sheeting attached to wooden frames (Supplement 1). As a control to account for potential “chamber effects” of plastic siding and the disturbance of burying heating cables, chicken wire was used for the sides of the final six plots and no resistance cables were buried within. For warming treatments, proximate chambers were randomly assigned within clusters of four-chamber ‘plots’ so that each block contained a non-chambered control plot (chicken wire sides), a control chamber with no warming (but with buried cables), a chamber with targeted warming 3°C above ambient and a chamber with targeted warming 5°C above ambient (n = 6 total replicates for each chamber/plot treatment; 24 chamber/plots in total). Each chamber and plot were outfitted for temperature and soil moisture monitoring and the real-time soil temperature data used to target the warming of treatment soils. During the months of the current study the real-time monitoring revealed that actual warming was ~2.5°C above ambient for all warming chambers (Supplement 2a), so we pooled the faunal data from these targets together. Given that there was also no difference in soil temperature between the non-chambered (chicken wire) control plots and the unheated control chambers (Supplement 2a), we also pooled faunal results from the non-chambered control plots and unheated control chambers. Overall, then, our design consisted of 12 warming chambers and 12 non-warmed ambient-temperature plots equally distributed between forest and edge habitats.
Deadwood assays
Two milled wood assays with no bark (12 x 14 x 2 cm), hereafter “assays,” were placed 1 m apart in each plot (n = 48 total) in April 2018 with half collected in June 2019 (n = 24) and half collected in June 2020 (n = 24) [Supplement 3]. The assays were made of Pinus taeda L. (loblolly pine), which is native and abundant in the study area. Each assay had a “G”-shaped chamber (width: 2.5 cm; depth: 2.5 cm) routed out and open to the outside to create potential nesting chambers for ant colonies, and each assay was placed in contact with the forest floor soil to promote termite foraging. Prior to placement in the field, the air-dry mass of each assay was recorded and ranged from 90 to 151 g, with a mean of 111 g. We note that the assays ranged in mass because of natural variation in wood density but all assays had the same dimensions, and the wood was not treated (e.g., chemically or by pressure) in any way. After weighing, each assay was topped with clear acrylic plastic and an opaque ceramic tile following established protocols from previous research in the area (Warren II et al. 2012, 2015b; Bradford et al. 2014). The clear top allowed periodic monitoring of the nests before collection, but the tile blocked light otherwise (because the study termites and ants are negatively phototropic).
We assessed ant abundance when the deadwood assays were collected by freeze-killing and counting ant abundance in each of the two years. Ants make their nests in the deadwood assays whereas termites channelize and feed on the wood. Termites in the study region are the only wood decomposers that import mineral soil into the wood [bonded with saliva and fecal material – often replacing the structure of consumed wood] (Ulyshen et al. 2014; Myer and Forschler 2019). We assessed termite activity by drying the emptied assays at 65°C for 7 d, weighing each assay, and then we calculated the ash-free dry mass by burning off the remaining wood at 500°C in a muffle furnace. Hence, termite activity was represented by the remaining mineral soil (g ∙ assay-1). This estimate likely better represents an estimate of integrated termite activity given that the termites forage in the assays, but nest underground. Termite abundance at the time of assay collection may only capture a snapshot of termite activity and a subset of colony workers. For example, from this study, 54% of the assays showed signs of termite activity (tunnels, soil filling) and live termites were found in 46% of the assays when collected in the first year; in the second year, however, 100% of the assays showed signs of termite activity, but termites were only found in 63% of the assays when collected. Ants and termites were identified by species with reference to voucher specimens and taxonomic guides. We note that ants and termites occurred naturally in the plots. Ants are the main predator of R. flavipes (Beard 1973; Warren II et al. 2015). Whereas many ant species occur at the study site (Warren II et al. 2022; Merchlinksky et al. 2023), few use wood nests to the degree of A. rudis and B. chinensis (we only found single colonies of Camponotus and Crematogaster spp. during the study).