Impacts of winter climate change on northern forest understory carbon dioxide exchange determined by reindeer grazing

Abstract
In northern regions, the ongoing climate change is altering snow depth with complex consequences for carbon dioxide (CO2) exchange and thus, global carbon (C) balance. In addition, ungulate grazers such as reindeer and caribou often alter plant and soil properties that may lead to modifications in the magnitudes and patterns of CO2 exchange. To understand how reindeer grazing, coupled with changes in snow depth affects CO2 exchange, we used recent snow treatments (ambient, reduced, and increased snow depth) combined with 25- and 55-year-old reindeer exclusions and the adjacent grazed areas in boreal and subarctic Scots pine forests that are main winter pastures for reindeer/caribou and cover a significant portion of boreal and subarctic landscapes. At both study sites, we measured understory net ecosystem exchange (i.e., NEE), ecosystem respiration (i.e., ER), and gross ecosystem production (i.e., GEP) over two snow-free seasons. We found that 55 years of reindeer exclusion increased C source strength by 140% under ambient snow depth and 205% under reduced snow depth in comparison to the grazed area with respective snow conditions. On the contrary, increased snow depth decreased C source strength inside the exclusion offsetting the difference between reindeer grazing treatments. Our results show that grazing may enhance ecosystem stability to winter climate change in comparison to long-term absence of grazing. This highlights the complexity of climate-grazer interactions in functioning of northern ecosystems which are experiencing variations in snow depth.
Methods
Experimental design
Study sites are located in two oligotrophic Scots pine (Pinus sylvestris L.) forests in northern boreal and subarctic Fennoscandia. Experimental setup of recent snow treatments crossed with reindeer (Rangifer tarandus L.) grazing treatments were used to study the impacts of changing snow depth and reindeer grazing on understory CO2 exchange.
At the boreal study site, the grazing treatment consisted of a 25-year-old reindeer exclusion and the adjacent grazed area whereas at the subarctic site, grazing treatment consisted of a 55-year-old reindeer exclusion and the adjacent grazed area. Study plots were established across these treatments: 18 plots at the boreal and 24 plots at the subarctic site. The study plots were organized into three blocks in the exclusion and in the grazed area at the boreal site (six blocks in total) and into four blocks in the exclusion and in the grazed area at the subarctic site (eight blocks in total). Each block contained three experimental plots assigned to three snow depth treatments including ambient, reduced, and increased snow depths. The snow treatments consisted of manually removing snow from the reduced snow plots to maintain a 25 cm snow depth throughout the winter, and the removed snow was added over the increased snow plots.
CO2 fluxes
CO2 fluxes were measured using manual chamber method on understory vegetation within exclusions and adjacent grazed plots across snow depth treatments (i.e., ambient, reduced, and increased snow depth). Measurements were conducted during the growing seasons of 2019, 2020 (boreal site), 2022 and 2023 (subarctic site) 1-4 times per week from June until October.
The measurement protocol consisted of two measurements at each study plot: net ecosystem exchange (NEE) under full light, then the chamber was vented, placed over the ground, and covered with an opaque chamber for measuring ecosystem respiration (ER). The system consisted of a custom-made chamber (2-mm thick transparent polycarbonate, diam. 30 cm × height 39 cm) equipped with a fan, a CO2 sensor (model GMP343, Vaisala Inc., Vantaa, Finland), and air temperature and humidity sensor (model HMP75, Vaisala Inc.). CO2 concentration (ppm), relative humidity (%), and temperature (℃) during each manual measurement were logged every 5 sec for 2 min with a handheld logger (model MI70, Vaisala Inc.). NEE and ER were calculated from the linear change of the CO2 concentration in the chamber by the ideal gas equation, and gross ecosystem productivity (GEP) was obtained by subtracting ER from NEE.
Soil temperature
Soil temperature was continuously measured at 5 cm depth with permanently installed sensors (T107 temperature probe, Campbell Scientific Inc., Logan, UT, USA) that were set to record soil temperature with 10-min interval (CR1000x data logger, Campbell Scientific Inc.) across snow and grazing treatments. Monthly average values were calculated per plot during respective study years. The data were obtained from Oulanka and Kevo research stations.
Snow depth
Snow depth was measured and mean snow depth for each plot was calculated approximately once a week and always immediately after snow treatments at the boreal site, and 1–2 times a month and always before snow treatments at the subarctic site. The snow treatments were carried out in 2018–2020 at the boreal site and in 2021–2023 at the subarctic site. The data were obtained from Oulanka and Kevo research stations.
Vegetation abundance
To map the impacts of exclusion on plant abundance, vegetation composition was recorded in July 2018 at the boreal site and in August 2020 at the subarctic site. At each study plot, the abundances of (reindeer) lichens, feather mosses (bryophytes), and vascular plants were recorded from 50 cm × 40 cm sized areas using a modified point-frequency method (Jonasson 1988) with ten pins and five rows (50 pins in total).