uid=HFR,o=lter,dc=ecoinformatics,dc=org all public read doi:10.6073/pasta/25d089b620c42e6aa53ab09a761298fa Israel Del Toro https://orcid.org/0000-0002-3901-8713 Aaron Ellison https://orcid.org/0000-0003-4151-6081 Relena Ribbons https://orcid.org/0000-0002-9566-3813 Researcher 2023 English

Direct and indirect consequences of global warming on ecosystem functions and processes mediated by invertebrates remain understudied but are likely to have major impacts on ecosystems in the future. Among animals, invertebrates are taxonomically diverse, responsive to temperature changes, and play major ecological roles which also respond to temperature changes. We used a mesocosm experiment to evaluate impacts of two warming treatments (+3.5 and + 5 °C, set points) and the presence and absence of the ant Formica subsericea (a major mediator of processes in north-temperate ecosystems) on decomposition rate, soil movement, soil respiration, and nitrogen availability. Replicate 19-Litre mesocosms were placed outdoors in lath houses and continuously warmed for 30 days in 2011 and 85 days in 2012. Warming treatments mimicked expected temperature increases for future climates in eastern North America. In both years, the amount of soil displaced and soil respiration increased in the warming and ant presence treatments (soil movement: 73 to 119%; soil respiration: 37 to 48% relative to the control treatments without ants). Decomposition rate and nitrogen availability tended to decrease in the warmest treatments (decomposition rate: -26 to -30%; nitrate availability: -11 to -42%). Path analyses indicated that ants had significant short term direct and indirect effects on the studied ecosystem processes. These results suggest that ants may be moving more soil and building deeper nests to escape increasing temperatures, but warming may also influence their direct and indirect effects on soil ecosystem processes.

ants climate change decomposition ecosystems nitrogen soil disturbance soil respiration LTER controlled vocabulary populations organic matter inorganic nutrients LTER core area Harvard Forest HFR LTER USA HFR default

This dataset is released to the public under Creative Commons CC0 1.0 (No Rights Reserved). Please keep the dataset creators informed of any plans to use the dataset. Consultation with the original investigators is strongly encouraged. Publications and data products that make use of the dataset should include proper acknowledgement.

Creative Commons Zero v1.0 Universal https://spdx.org/licenses/CC0-1.0.html CC0-1.0 https://harvardforest.fas.harvard.edu/exist/apps/datasets/showData.html?id=hf244 Harvard Forest Experimental Lath Houses. Coordinates based on WGS84 datum. -72.19 -72.19 +42.53 +42.53 330 330 meter 2011 2012 genus Formica species subsericea complete Information Manager Harvard Forest

324 North Main Street Petersham MA 01366 USA

(978) 724-3302 hf-im@lists.fas.harvard.edu Harvard Forest

324 North Main Street Petersham MA 01366 USA

(978) 724-3302 (978) 724-3595 https://harvardforest.fas.harvard.edu

Mesocosm experiments have been used extensively to test hypotheses derived from field observations that cannot be tested with large-scale field experiments because of logistic constraints (Sala et al. 2000), and can help simplify complex systems to identify the impact of individual manipulated variables on a constrained number of measured responses. Our mesocosm experiment was a two-way factorial experimental design with two levels of warming (+3.5 °C and +5 °C set-points) crossed with three treatments of soil manipulation: 1) addition of ants (Formica subsericea: N = 96 [2011]; N = 82 [2012]); 2) undisturbed controls (N = 15 [2011]; N = 24 [2012]; and 3) a disturbance control, in which every week, the upper 3-5 cm of the top soil layer was manually overturned the with a trowel and the bottom soil layers were aerated with a 25-cm pick (N = 12 [2012 only]). The disturbance control was established to evaluate the effects of manual soil manipulation in comparison to soil manipulation done by ants on the measured response variables. The experimental warming temperatures that we used represented expected summer temperature increases for eastern North America by 2100 under two different climatic change scenarios (i.e. +3.5 ˚C, RCP6.0 scenario and +5.0 ˚C, RCP8.5 scenario; Deser et al. 2012; IPCC 2013). The experiment ran for 30 days in 2011 (June − July) and 85 days in 2012 (May − July). Projected increases in temperatures are likely to reach the critical thermal maxima (Oberg, Del Toro and Pelini 2012) of some common ant species of eastern North America (including F. subsericea) that can ultimately have negative impacts on the ant physiology and functionality of ants in temperate ecosystems. Each mesocosm was contained in a 19 L plastic bucket (30 cm diameter and 39 cm height). The containers were filled with three layers of soil, which mimicked the soils where F. subsericea commonly occurs. The lowest layer consisted of 2.0 L of 1.2-cm sized gravel and 2.0 L of sand; the middle layer consisted of 5.5 L of sand and 2.0 L of gravel; and the top layer consisted of 2.0 L of sand and 2.0 L of heat treated compost. We drilled eight 1-mm diameter holes at the base of each mesocosm to allow excess water to drain out. The top 2 cm of each container was coated with petroleum jelly to prevent the ants from escaping each mesocosm. We cut a 14-cm diameter hole on each of the mesocosm lids and covered the opening with 1-mm fiberglass mesh. The mesh allowed ambient humidity and the heat from the lamps to penetrate the mesocosm, while excluding potential predators and seeds and preventing the ants from leaving the mesocosm. Temperature sensors were placed in the middle of each soil layer, by drilling 5-mm diameter holes and inserting thermocouples into the mesocosm. The holes were then completely sealed with silicone. Mean, minimum, and maximum hourly temperatures were recorded on a CR1000 data-logger (Campbell Scientific, Logan, Utah, USA) and downloaded weekly. Humidity was recorded using a CD620 HydroSense water-content sensor (Campbell Scientific) and monitored three times per week to make sure that all treatments remained at a constant soil moisture (5 – 8% soil moisture content). If soil moisture in the mesocosms fell below 5%, they were watered for 5 minutes with a sprinkler irrigation system. Such an extreme reduction in soil moisture occurred only three times in the +5 °C warming treatment and once in the +3.5 °C warming treatment throughout the duration of the experiment. The mesocosms were actively warmed using 250-Watt infra-red heat lamps (Phillips 250W, 120 volt heat light). We chose infra-red lighting because we believe this color spectrum would not affect diurnal and nocturnal activity and behaviors in the ants. Previous work with the ant Formica cuniculaia (a closely related species of F. subsericea in the fusca group) shows this species to have a dichromatic visual system and is unable to distinguish color differences at wavelengths greather than 540 nm (Aksoy and Camlitepe 2012) therefore we believe the infrared lightling was not detectable by the ants but the temperature manipulation was. To achieve a mean temperature increase of 3.5 °C, the lamps were hung 120 cm above the surface of the mesocosms. To achieve a mean temperature increase of 5 °C, the lamps were hung 60 cm above the surface of the mesocosms. Lamp bulbs were left on continuously so that daily temperature fluctuations were consistent across the treatments; bulbs were replaced as needed. One lamp evenly heated four mesocosms through all soil layers, so we clustered the mesocosms in groups of four; the distribution of the soil manipulation treatments was randomized within each temperature treatment. Actual hourly mean temperatures experienced within each mesocosm in the “+3.5 °C treatment” ranged from 0.5 to 3.9 °C warmer than controls, whereas the “+5 °C treatment” ranged from 1.7 to 6.7 °C warmer than controls, with the largest temperature differences occurring during the coolest and warmest times of the day. These soil temperatures are consistent with air and soil temperature variation observed at another Harvard Forest field site where open-top chambers are being warmed (Pelini et al. 2011) to replicate IPCC warming projections and at the Fisher meteorological station in Harvard Forest which documents daily temperature variation.

Formica subsericea is a common and widely distributed species that builds large nests in soil. Its large colonies (hundreds to > 10,000 workers) typically are found in edge and open habitats throughout eastern North America (Ellison et al. 2012). The genus Formica has a Holarctic distribution and many of the species in the diverse and widespread fusca group (which includes F. subsericea) share similar natural history traits (Francoeur 1973; Gösswald 1989). We expect, therefore, that the results from our experiment are likely to apply to other north temperate-zone soils where Formica fusca-group species occur. We collected ant colonies from sandy soils at three localities in central Massachusetts: the Montague Pine Barrens (42.569 ˚N, -72.536 ˚W), the Devens Pine Barrens (42.452 ˚N, -71.641 ˚W), and Myles Standish State Forest (41.839 ˚N, -70.691 ˚W). Each mesocosm was established with 100 workers from independent ant colonies; no queens were collected or used in the mesocosms. We expected that individual activity and behavior would not be affected by the absence of queens as ants self-organize tasks without the need of direction from a reproductive caste (Oster and Wilson 1978; Gordon 2010). To keep the ants alive for the duration of the experiment, the ants in each mesocosm were fed with a 5% sugar and 5% glutamine solution, which mimics honeydew, stored in a 50-ml vial and replaced on a weekly basis. The aqueous solution was contained in the vials with a cotton ball and did not in any way leach into the soil filling the mesocosm. We counted and removed the number of dead workers in each mesocosm on a weekly basis. Mortality rates ranged from 10-22 ants per mesocosm and there were no significant differences in mortality rates across treatments or between years.

We measured mass loss (as an indication of decomposition rate) by placing two fiberglass mesh bags filled with leaf litter on top of the soil in each mesocosm. Each bag contained 1 g of dried (70 °C for 72 hr) whole red maple (Acer rubrum L.) leaves + 1 g of dried whole red oak (Quercus rubra L.) leaves. One litter bag excluded ants because of the size of the mesh (1-mm), while the second bag allowed ants access to the litter (mesh size = 10-mm). At the end of the experiment, litter bags were removed from the mesocosms; the remaining litter was dried (70 °C for 72 hr) and weighed (± 0.001g), to determine mass lost. At the end of the experiment, the 10-mm mesh bags contained much smaller portions of leaves and often had ants still inside the bags when opened (personal observation), whereas no ants were ever detected in the 1-mm mesh bags. Soil respiration (as CO2 flux) was measured for 5 min every 7 days in 2011 and every 14 days in 2012 using a portable infra-red gas analyzer (LI-COR 6400, LI-COR Biosciences, Lincoln, Nebraska) with a soil-flux chamber (LI-COR 6400-09) placed on a PVC collar (20-cm diameter, 5-cm deep) half-buried at the soil surface in a subset of the mesocosms (N = 38 [2011]; N = 72 [2012]). We used linear interpolation and integration (based on the relationship between measured soil temperature and CO2 flux; Savage, Davidson and Tang 2013) to estimate the amount of CO2 produced per day over the course of the experiment each year. We determined the total amount of available nitrogen (NH4+ and NO3-) as that captured on ion-exchange resin bags placed in each mesocosm (Maynard, Kalra and Crumbaugh 2008). Approximately 5 g of resin was placed in a nylon mesh bag and pretreated with 2 mol L-1 KCl before being placed in the mesocosms. Resin bags were placed 3 cm below the surface of the soil and left in the mesocosm for the duration of the experiment during the 2012 season. Resins were returned to the laboratory and dried at 105˚C for 24 hours. Resins were then extracted in 100 mL of 1 mol L-1 KCl for 48 hours. Resin extracts were filtered through a coarse pore filter (0.45-0.60 µm), and inorganic N concentrations were determined colorimetrically with a Lachat AE flow-injection auto-analyzer (Lachat Instruments, Inc., Loveland, Colorado, USA) using the indophenol-blue and cadmium reduction methods for NH4, and NO3 respectively. Nitrogen mineralization is a reliable predictor of soil productivity (Raison, Connell and Khanna 1987), and total nitrogen captured on resins is closely correlated with nitrogen mineralization (Strader, Binkley and Wells 1989). Net nitrogen capture was determined by subtracting the concentration of inorganic nitrogen (NH4+ + NO3-) in the incubated samples from that of the initial samples (Morecroft, Marrs and Woodward 1992). High net nitrogen capture can be indicative of high levels of microbial activity in the soil (Chapin, Matson and Mooney 2002). We estimated the amount of soil moved by F. subsericea by sacrificing the colony at the end of the experiment and creating a three-dimensional paraffin cast of the nest (Gulf Wax®, Royal Oak, Roswell, GA) as described by Tschinkle (2010). We estimated the volume of soil displaced from the weight of the cast and the density of the paraffin (0.9 g/cm3). We also recorded the maximum depth (cm from the surface) for every cast.

The zip file (hf244-06) contains R code and output for two analyses: (1) ANOVA tables and pairwise comparisons for decomposition, nitrogen, soil movement, and soil respiration data (hf244-01 through hf244-04), and (2) Pathmodel analysis for path anaylsis data (hf244-05).

Harvard Forest

324 North Main Street Petersham MA 01366 USA

(978) 724-3302 (978) 724-3595 https://harvardforest.fas.harvard.edu https://ror.org/059cpzx98 pointOfContact The Harvard Forest Long-Term Ecological Research (LTER) program examines ecological dynamics in the New England region resulting from natural disturbances, environmental change, and human impacts. National Science Foundation LTER grants: DEB-8811764, DEB-9411975, DEB-0080592, DEB-0620443, DEB-1237491, DEB-1832210. hf244-01-decomposition.csv decomposition raw data hf244-01-decomposition.csv 12878 ba4befb2f78132e48276b92a8a439166 1 \r\n column , https://harvardforest.fas.harvard.edu/data/p24/hf244/hf244-01-decomposition.csv year year of data collection YYYY 1 year NA missing value id mesocosm identifier mesocosm identifier NA missing value treatment experimental warming treatment Control control Plus3.5 set-point of +3.5 °C Plus5 set-point of +5 °C NA missing value treatment.2 experimental soil manipulation treatment Ants addition of ants Empty undisturbed control Manual disturbance control NA missing value bag mesh size of decomposition bag mesh size of decomposition bag NA missing value mass.lost dry weight of leaf litter mass lost from the corresponding decomposition bag gram 0.001 real NA missing value 340 hf244-02-nitrogen.csv nitrogen raw data hf244-02-nitrogen.csv 6635 0972e2681ac634f2306a9ea966b49a01 1 \r\n column , https://harvardforest.fas.harvard.edu/data/p24/hf244/hf244-02-nitrogen.csv year year of data collection YYYY 1 year NA missing value id mesocosm identifier mesocosm identifier NA missing value treatment experimental warming treatment Control control Plus3.5 set-point of +3.5 °C Plus5 set-point of +5 °C treatment.2 experimental soil manipulation treatment Ants addition of ants Empty undisturbed control Manual disturbance control NA missing value no2.no3 total nitrate obtained from resin bags milligramsPerLiter 0.0001 real NA missing value nh3 total ammonia obtained from resin bags milligramsPerLiter 0.001 real NA missing value no3 nitrate milligramsPerLiter 0.00001 real NA missing value available.n total amount of available nitrogen (NH4+ and NO3-) determined as that captured on ion-exchange resin bags placed in each mesocosm milligramsPerLiter 0.00001 real NA missing value mineralization.rate amount of nitrogen captured in the resins over the duration of the experiment calculated as the total amount of ammonium (NH4) and nitrate (NO3) captured in mg N/L/day (based on the number of days the resin bags were deployed in the experiment) milligramsPerLiterPerDay 0.000000001 real NA missing value nitrification.rate amount of nitrate (NO3) produced and captured in the resin bags in mg N/L/day (based on the number of days the resin bags were deployed in the experiment). milligramsPerLiterPerDay 0.000000001 real NA missing value 106 hf244-03-soil-movement.csv soil movement raw data hf244-03-soil-movement.csv 3964 9f912a0bf263073c537f5e6a0a166187 1 \r\n column , https://harvardforest.fas.harvard.edu/data/p24/hf244/hf244-03-soil-movement.csv year year of data collection YYYY 1 year NA missing value id mesocosm identifier mesocosm identifier NA missing value treatment experimental warming treatment Control control Plus3.5 set-point of +3.5 °C Plus5 set-point of +5 °C NA missing value treatment.2 experimental soil manipulation treatment Ants addition of ants NA missing value mean.temp mean temperature celsius 0.00000001 real NA missing value total.volume otal volume of nest cast gram 0.0001 real NA missing value no.entrances number of entrances number 1 whole NA missing values max.depth maximum depth of the deepest chamber or tunnel in each mesocosm centimeter 1 whole NA missing value largest.depth depth of largest chamber centimeter 0.1 real NA missing value no.chambers number of chambers centimeter 1 whole NA missing value no.tunnels number of tunnels centimeter 1 whole NA missing value longest.length length of longest tunnel centimeter 0.1 real NA missing value 66 hf244-04-soil-respiration.csv soil respiration raw data hf244-04-soil-respiration.csv 4956 1a5c1ebc93ffc190fbe21212744a2042 1 \r\n column , https://harvardforest.fas.harvard.edu/data/p24/hf244/hf244-04-soil-respiration.csv year year YYYY 1 year NA missing value id mesocosm identifier mesocosm identifier NA missing value treatment mesocosm identifier Control control Plus3.5 set-point of +3.5 °C Plus5 set-point of +5 °C NA missing value treatment.2 experimental soil manipulation treatment Ants addition of ants Empty undisturbed control Manual disturbance control NA missing value total.co2 co2 production in one hour in ųmol/m-2, calculated based on 5 minute test interval micromolePerMeterSquaredPerSecond 0.00001 real NA missing value daily.flux estimated daily CO2 production in ųmol/m-2 micromolePerMeterSquaredPerDay 0.0001 real NA missing value 102 hf244-05-path-analysis.csv path analysis raw data hf244-05-path-analysis.csv 5220 ae053854eb092d9dcd4162040df87f17 1 \r\n column , https://harvardforest.fas.harvard.edu/data/p24/hf244/hf244-05-path-analysis.csv year year of data collection YYYY 1 year NA missing value id mesocosm identifier mesocosm identifier NA missing value warming experimental warming treatment (average degrees C of warming above the control treatment) celsius 0.01 real NA missing value ants presence or absence of ants in mesocosms 0 absence of ants 1 presence of ants NA missing value empty empty or occupied mesocosm 0 occupied mesocosm 1 empty mesocosm NA missing value soil.displaced total volume of nest cast in grams as volume of soil displaced gram 0.0001 real NA missing value mass.lost.10 leaf litter mass lost from 10mm mesh bags gram 0.001 real NA missing value mass.lost.01 leaf litter mass lost from 1mm mesh bags gram 0.000001 real NA missing value daily.co2 estimated daily CO2 production in ųmol/m-2 micromolePerMeterSquaredPerDay 0.000001 real NA missing value no3 total nitrate obtained from resin bags milligramsPerLiter 0.00001 real NA missing value nh3 total ammonia obtained from resin bags milligramsPerLiter 0.001 real NA missing value 93 hf244-06-R-code.zip R code hf244-06-R-code.zip 259645 ec012ab28f1ba19c209436700d6de0da zip html https://harvardforest.fas.harvard.edu/data/p24/hf244/hf244-06-R-code.zip script soil short-term measurement milligramsPerLiterPerDay micromolePerMeterSquaredPerDay