<scp>SWAMP</scp>: A new experiment for simulating permafrost warming and active layer deepening on the Tibetan Plateau

Permafrost (ground that keeps frozen at or below 0°C for two more consecutive years; Subcommittee, 1988) covers about 2.3 × 107 km2, accounting 16% of the global land area (Zhang et al., 2008). In Northern Hemisphere, permafrost is primarily found high-latitude region, example Alaska and Siberia, in high-altitude areas, Tibetan Plateau (Obu 2019). these regions, microbial decomposition limited due to prolonged low temperature freezing conditions, which turn leads an accumulation soil organic carbon (SOC; Johnston The pool top 3 m across regions has recently been estimated be 1014 Pg C (1 = 1015 g), approximately double size atmospheric (Mishra 2021). However, this large currently threatened by continuous climate warming. particular, with zero annual amplitude (the depth yearly fluctuations fall 0.1°C; French & Shur, 2010) experienced a significant warming 0.29°C over last decade (Biskaborn rise accelerates thaw, resulting increase active layer thickness (e.g. 0.15 cm per year Arctic 1.95 on Plateau; Wang 2022) abrupt thaw formation thermokarst landscape features such as lake; Olefeldt 2016). These changes thermal state deposits may leave SOC available decomposition, subsequently release quantities greenhouse gases [e.g. dioxide (CO2) methane (CH4)], potentially create positive feedback between (Johnston 2019; Miner 2022; Schuur 2015). there considerable variation model prediction under impacts change [ranging from −641 167 Representative Concentration Pathway (RCP) 8.5, highest baseline emission scenario throughout 21st century; McGuire 2018]. To minimize uncertainties, various approaches have used situ experiments (Henry Molau, 1997), observations gradients (Hugelius Kuhry, 2009), laboratory incubation (Chen Manipulative important place among since they overcome most limitations gradient studies (Torn Therefore, synchronous simulations subsequent deepening can greatly contribute comprehensive understanding carbon-climate feedback, especially regarding direction magnitude feedback. During past several decades, great efforts made better understand responses ecosystems multiple manipulative 1997; Natali 2011). best known International Tundra Experiment (ITEX), established tundra 1990. Using open chambers (OTCs) simulate warming, ITEX explored phenology, growth functional traits vascular plants response 1997). Although OTCs are effective air surface soil, effects deep >50 cm) very (Marion constrains our dynamics thaw. With limitation mind, Carbon Experimental Heating Research (CiPEHR) combined both ITEX-style snow fences make successful strides simulating (Natali passive techniques where shallow ~50 Arctic; Luo 2016), but their inadequate when thick than 100 2022). urgent need develop novel approach achieve ecosystem layer, simulation well deepening. development next-generation whole-soil-profile whole-ecosystem technologies offers opportunity thawing, thickness. utilize simultaneously elevate entire profile (Hanson technologies, 1–2 m) increased 4–9°C 2017; Hicks Pries Nottingham 2020; Qin 2023). aims establish realistic future explore structure function mechanisms involved For ecosystems, processes particularly because vulnerable consequently profound balance (Koven As yet, however, no reports application modified versions, remains constrained experimental limitation. Here, we introduce experiment, Simulate Warming Mountain (SWAMP), 2021 swamp meadow Plateau. Combining concepts SWAMP implemented redesigned system effectively simulates during growing season, Infrared heaters were warm heating rods being inserted vertically employed transmit heat into and, deposits. characterize SWAMP, measurements whole (from table) moisture depths taken full year. addition, thawing duration different also recorded May October. main objective present design performance first experiment utilizing technique. unique seek provide shed insights how particular dynamics, will respond hosts 1.1 106 km2 largest world middle latitudes), accounts 40.2% region (Cheng mean precipitation range −4.9 6.1°C 84.3 593.9 mm, respectively (Ding plateau contains variety alpine including forest, shrubland, meadow, steppe desert, grassland predominating 1988). Large amounts stored stock upper 15.3–21.7 2016; Wang, Wu, Over three average (~0.5°C decade; Kuang Jiao, 2016) twice rate same period (0.3–0.5°C 0–40 [see Figures S1] 0.02–0.3°C 10 depth; Wu Zhang, Under impact increasing 5–20 decades (Peng 2018; Current give values 100–320 (Xu scenarios, predicted 6–8 (Shared Socioeconomic Pathways [SSP] 245; medium pathway gas emissions) 18–33 (SSP 585; boundary towards end century 2021; Zhang located east Altun-Qilian ranges (37.50°N, 100.24°E; elevation: 3670 a.s.l.; Figure 1a), one five typical (Jin site characterized highland continental climate, warm, humid summer cold, dry winter (Li 2020). 1959–2020 was −0.1°C, varied 0.3°C (Figure S3). monthly study fluctuates minimum −13.3°C January maximum 11.2°C July (data National Meteorological Information Center China). Mean 364.6 92% fell September within sporadic 120 Permission fieldwork non-required. vegetation dominant types (alpine steppe, meadow) density (carbon amount area) plant community dominated sedge, Carex tibetikobresia, kokanica atrofusca, species making up nearly 78% cover. Other Triglochin maritima, scabrirostris Pedicularis longiflora etc. (Table S1). Vegetation coverage 100%, richness 11.2 ± 0.8 m−2 (this data given SE) above-ground biomass 210.0 13.0 g m−2. classified FAO Cambisols nutrient content (0–10 extremely high, 209.6 2.8 mg g−1, total nitrogen 16.2 0.2 g−1 dissolved inorganic (DIN) 21.6 1.7 kg−1. pH 6.4 0.1, value associated (0.28 0.03 g−1). Soil carbon, other physicochemical properties not significantly plots selected control any (Figures S4–S6 Table S3), indicating little spatial heterogeneity area. DIN comparable S5), line observed patterns (Beermann Salmon 2018). 2021. fenced avoid anthropogenic disturbance, set gently sloping terrain 50 1b). paired consisted six replicated blocks eliminate influence topography microenvironment. Each replicate block contained treatments: plot circle diameter 7.1 m2, buffer zone ~3 neighbouring plots. total, occupied 600 m2 2a). heater (Kimball, 2005) while 2011) provided below-ground 3). actively permafrost). 2°C ~2°C RCP 4.5 (an intermediate Soong 2020), raise compared 2020 Above ground, infrared (8 mm 151 long, 2000 W, 240 V Model, HS-2420, Kalglo Electronics, USA) suspended parallel height 1.5 above ground homogeneous possible without overheating (Kimball When operation, emitted radiation warmed soil. part, 21 (20 around perimeter circular equal spacing center; steel tube) installed 2b). cables (10 width, 33 W m−1, Mode, BSX 10-2-FOJ, Thermon, each rod, surrounding void filled quartz sand hold cable enhance efficiency transfer 80 long overall, 60 buried remaining 20 exposed ground. above- systems controlled separate devices, workflows almost identical. general, collected using logger; secondly, difference pair treatments calculated via computation module; thirdly, devices governed based Specifically, system, thermocouple probes (WZP-231, Xincheng Electric Appliance, China; Range: −50~250°C; Accuracy: ±0.3°C) (0–5 acted sensors heaters. arranged heaters, placed equidistant centre edge 2b Temperature module (Current/Power Output Module, Haiwell Happy, activated less warmer those plot. They turned off exceeded 2°C. T-type thermocouples (TX-FFRP, Shanghai Kaixin Sensor Technology, −40~200°C; cavity positioned adjacent seven (10, 20, 40, 60, 80, [from table]). Data logger (CR1000, Campbell Scientific). 40 plot, on. greater 2°C, Dummy mimic thickness, simulated temperatures air, season (May–September). reasons, actions undertaken outside season. First, kept (soil > 0°C; S8). fact, current scenario, stays taliks < S8), so would result excessive unrealistic scenario. Second, primary aim process occurs mainly 2021, late April (27th April) its trial run ended August (29th August), having running 124 days. time, procedures equipment constantly tested adjusted. Based (i.e. thickness), second-year initiated 27th April, 2022, began ended, after 160 days, 4th October, freeze. Air temperature, continuously monitor environmental factors. An probe (DS1923 iButtons, Maxim Integrated) measurement hole eight table cm). Considering high water flooding 2022), fact damaged if too deeply, volumetric measured four cm; 5TM-EM50, Decagon). All 1-hour intervals. Altogether, probe, Thawing thaw-probe times month thawing. Four points equally determine (period nonfreezing), start time manually thaw-probe, day onwards (Wang First all, one-hour moisture) averaged daily scale. evaluate depth, linear mixed effect models treatment fixed variable, nested observation date (day) random factors account repeated measurements. Then, examine thickness) depths, applied variable. homogeneity plots, samples before 2020) content, pH; details variables, see Appendix We performed properties, division control) did normally distributed residuals log- arcsin-transformed. p adjusted ‘Bonferroni’ method (Benjamini Hochberg, 1995). statistical analyses conducted R v4.1.2 (R Core Team, lme4 lmerTest packages perform (Kuznetsova 2017). Daily seasonally, ranging −13.7 0.9°C results 0.2°C 4a). exhibited strong seasonal variation, lowest non-growing (0–100 February March) (0–60 August, 80–120 September), differences seasons decreased 5). operational 4 5 S2). 1.8 0.6°C, 11.4 0.1°C 13.3 0.6°C 4b). 5a–e), increases 2.1 cm), cm, 1.9 2.0 0.4°C 5i). Below 5f–h), 1.4 0.6 contrasting variations temperature. remained level, 51.1 1.7% 47.6 1.2% 37.1 1.4% 21.4 1.6% 6a–d). Similar affected 6a–d had minor it 1.8%, 52.9 1.0% 6a). grew gradually 4.3 2.1% 6.0 3.4% 12.3 2.3% 6e). completed early accelerated thawed 7). 7a). occurred (33.2 2.4 37.6% promoted 10.5% occurring October (control: 120.0 warmed: 132.6 12.6 7b). 7c). S2), smallest 22.8 3.3 days 49.3 cm. Our showed steadily elevated This consistently target insertion whole-soil-warming experiments, lower subsoil. probably dissipation supply (Hicks By combining compensate insufficient warmth overcame shortcoming achieved Moreover, pattern largely long-term (Hu layers identical (increased 0.44–0.45°C weakened (0.24–0.40°C 200 few induced layers, rather direct column declined (Liu According obtained boreholes (~0.2°C decade) roughly (0.09–0.12°C temporal cold seems consistent trend similar (Soong possibly soils specific capacity higher (Yadav Saxena, 1973), intake become latent sensible (Halliwell Rouse, 1987). therefore provides relatively trends gradual decrease deeper layers. determines extent degree Before many attempts appropriate comparing (more (More permafrost, rapid degradation collapse creating lakes S9a–c). climatic prevalent form gets warmer, landscapes might form; period, Within next years, heat, anticipated simulated. received only modest S8g,n), collapsed S9i). appropriately collapsing rapidly, S8g,n S9h). rod ultimately directly down Such allowed reported occur widely utilized designs Rather ‘multi-loop rod’ layout (heating loops plot; Hanson S10a) ‘single-loop + multi-loop cable’ (one loop inside 2023; S10b) ‘loop point S10d). centre. ‘20 1’ but, according desired effect. Compare central energy consumption, suitable small size. addition differences, desig