This is the accepted version of the journal article: Li, Jiwei; Deng, Lei; Peñuelas, Josep; [et al.]. «C:N:P stoichiometry of plants, soils, and microorganisms : response to altered precipitation». Global change biology, Vol. 29, issue 24 (Dec. 2023), p. 7051-7071. DOI 10.1111/gcb.16959 This version is available at https://ddd.uab.cat/record/284786 under the terms of the license 1 C:N:P stoichiometry of plants, soils, and microorganisms: Response to altered 2 precipitation 3 Jiwei Lia,b, Lei Denga,b*, Josep Peñuelasc,d, Jianzhao Wua, Zhouping Shangguana,b, Jordi 4 Sardansc,d, Changhui Penge, Yakov Kuzyakovf,g 5 a 6 State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau, Institute of 7 Soil and Water Conservation, Northwest A&F University, Yangling, Shaanxi 712100, China b 8 Institute of Soil and Water Conservation, Chinese Academy of Sciences and Ministry of Water 9 Resources, Yangling, Shaanxi 712100, China c 10 CREAF, Cerdanyola del Vallès, Barcelona, Catalonia, Spain d 11 CSIC, Global Ecology Unit CREAF-CSIC-UAB, Bellaterra, Barcelona, Catalonia, Spain e 12 Center of CEF/ESCER, Department of Biological Science, University of Quebec at Montreal, 13 Montreal, H3C 3P8, Canada f 14 Department of Soil Science of Temperate Ecosystems, Department of Agricultural Soil 15 Science, University of Goettingen, Gottingen, Germany g 16 Peoples Friendship University of Russia (RUDN University), 117198 Moscow, Russia 17 18 * Corresponding author: Prof. Lei Deng 19 Address: No. 26 Xinong Road, Yangling, Shaanxi 712100, China 20 Phone: +86-29-87012884, Fax: +86-29-87012210 21 Email: leideng@ms.iswc.ac.cn 22 1 23 Abstract Precipitation changes modify C, N, and P cycles, which regulate the functions 24 and structure of terrestrial ecosystems. Although altered precipitation affects above- 25 and belowground C:N:P stoichiometry, considerable uncertainties remain regarding 26 plant–microbial nutrient allocation strategies under increased (IPPT) and decreased 27 (DPPT) precipitation. We meta-analyzed 827 observations from 235 field studies to 28 investigate the effects of IPPT and DPPT on the C:N:P stoichiometry of plants, soils, 29 and microorganisms. DPPT reduced leaf C:N ratio, but increased the leaf and root N:P 30 ratios reflecting stronger decrease of P compared with N mobility in soil under drought. 31 IPPT increased microbial biomass C (+13%), N (+15%), P (26%), and the C:N ratio, 32 whereas DPPT decreased microbial biomass N (−12%) and the N:P ratio. The C:N and 33 N:P ratios of plant leaves were more sensitive to medium DPPT than to IPPT because 34 drought increased plant N content, particularly in humid areas. The responses of plant 35 and soil C:N:P stoichiometry to altered precipitation did not fit the double asymmetry 36 model with a positive asymmetry under IPPT and a negative asymmetry under extreme 37 DPPT. Soil microorganisms were more sensitive to IPPT than to DPPT, but they were 38 more sensitive to extreme DPPT than extreme IPPT, consistent with the double 39 asymmetry model. Soil microorganisms maintained stoichiometric homeostasis, 40 whereas N:P ratios of plants follow that of the soils under altered precipitation. In 41 conclusion, specific N allocation strategies of plants and microbial communities as well 42 as N and P availability in soil critically mediate C:N:P stoichiometry by altered 43 precipitation that need to be considered by prediction of ecosystem functions and C 44 cycling under future climate change scenarios. 45 KEY WORDS 46 Climate zone; Ecological stoichiometry; Precipitation change; Plant-soil interactions; Soil 47 microbes; Terrestrial ecosystems; Global change 2 48 1 | INTRODUCTION 49 Global warming and extreme weather events have large impacts on global precipitation patterns 50 (Maurer et al., 2020; Smith et al., 2023). Precipitation changes can shift the magnitude and 51 frequency of extreme precipitation events (Deng et al., 2021; Stocker, 2014), and affect plant 52 growth, litter production quality and quantity. Precipitation changes can also alter soil pH, 53 aggregate stability, and cation exchange capacity, and consequently modify plant and microbial 54 community stability (Austin & Vitousek, 2000; Khalili et al., 2016; Ma et al., 2016). These 55 changes have a large influence on carbon (C), nitrogen (N), and phosphorus (P) cycles, as well 56 as their stoichiometry (Prommer et al., 2019; Sierra et al., 2017; Thakur et al., 2015). Therefore, 57 clarifying the changes of C:N:P stoichiometry to altered precipitation is essential to predict the 58 impacts of climate change on ecosystem functions. 59 Ecological stoichiometry has been widely used to study the balance of energy and nutrients 60 at the level of individual organisms, communities and ecosystems (DeMott & Pape, 2005). 61 Changes in stoichiometry of nutrients reflect the ecosystem limitations and consequently, the 62 response to any environmental factors (e.g. N or P deposition), and also to future global change 63 (Sardans et al., 2012; Yue et al., 2017). Ecosystem C:N:P stoichiometry mainly affect the 64 primary productivity, nutrient cycling and food webs in ecosystems (McGroddy et al., 2004). 65 The C:N:P stoichiometry are related to litter decomposition, N2 fixation and leaching, plant 66 diversity, and the ability of organisms to adapt to environmental stresses (Heuck et al., 2015; 67 Nielsen & Ball, 2015). An asymmetric model of the relationship between the aboveground net 68 primary production (ANPP) and altered precipitation (Knapp et al. 2017) indicated positive and 69 negative asymmetry responses of ANPP to the precipitation variability (Knapp et al., 2017). 70 The positive asymmetry indicates that the ANPP change to increased precipitation (IPPT) is 71 higher than to decreased precipitation (DPPT). The negative asymmetry indicates that the 72 negative impact of extreme DPPT on ANPP is far greater than the positive effects of extreme 3 73 IPPT (Knapp et al., 2015; Zhou et al., 2018). The asymmetric reactions of terrestrial C:N:P 74 stoichiometry to IPPT or DPPT, however, are still unknown, which limit our predictions on the 75 C balance of terrestrial ecosystems to precipitation changes. 76 The results of altered precipitation on plant, soil and microbial C:N:P stoichiometry are 77 inconsistent (Sardans et al., 2021). For example, an increase in plant biomass N:P ratio by DPPT 78 due to a larger decrease in plant P concentration than N concentration (He and Dijkstra, 2014). 79 A meta-analysis generalized that DPPT increases plant N:P ratios (Yuan & Chen, 2015), owing 80 to that P is far more immobile than N, and drought decrease the uptake of P more than that of 81 N (Sardans et al., 2021). Two meta-analyses have demonstrated contrasting results on soil 82 microorganisms, i.e., one reported microbial biomass C:N ratio decreased with increased 83 precipitation (Xu et al., 2020), but the other reported the increase of C:N ratio (Deng et al., 84 2021). These inconsistences were attributed to: (1) magnitude of precipitation manipulation 85 (Ren et al., 2017; Wilcox et al., 2017); (2) duration of precipitation manipulation (Canarini et 86 al., 2017; Zhou et al., 2018); (3) differences in local climate [e.g., mean annual temperature 87 (MAT) and mean annual precipitation (MAP)] (Wu et al., 2011; Xu et al., 2020). These studies, 88 however, have mainly focused on the effects of altered precipitation on a single indicator (e.g., 89 C, N, or C:N) of plant, soil or microbial C:N:P stoichiometry. The comprehensive effects of 90 precipitation changes on plant-soil-microbial system were rarely studied. Furthermore, the 91 extent of the magnitude, duration, and climate factors on plant, soil and microbial C:N:P 92 stoichiometry is still unclear. 93 To generalize the response of plant-soil-microbial C:N:P stoichiometry to altered 94 precipitation, we conducted a synthesis from precipitation manipulation experiments in three 95 groups of ecosystems: grasslands, shrublands and forests. We meta-analyzed 828 observations 96 from 236 studies to answer the following questions: (1) how do C:N, C:P, N:P ratios of plants, 97 soil and microorganisms respond to IPPT and DPPT? (2) How do these responses and 4 98 sensitivity depend on climate and ecosystem types? and (3) Which factors affect the response 99 of C:N:P stoichiometry to altered precipitation? 100 101 2 | MATERIALS AND METHODS 102 2.1 | Data compilation and extraction 103 A comprehensive search of primary studies with experimental precipitation manipulation (i.e., 104 increased precipitation and/or drought) published before MAY 2022 were selected. These 105 studies evaluated the effects of increased (IPPT) and decreased (DPPT) precipitation on C:N:P 106 ratios in plants, soils, and microbial biomass. We obtained the published papers from Google 107 Scholar (www.scholar.google.com), Web of Science (www.webofscience.com), and China 108 Knowledge Infrastructure (CNKI; www.cnki.net). The following search terms were used: 109 (“altering precipitation” OR “changing precipitation” OR “altering rainfall” OR “changing 110 rainfall” OR “drought” OR “decreasing precipitation” OR “increasing precipitation” OR 111 “precipitation variation” OR “water addition” OR “water reduction” OR “precipitation 112 experiment”) and (“C” OR “N”, “P”, OR “C:N” “C:P” OR “N:P” OR “C:N:P”). We filtrated 113 the published papers using the following criteria in the dataset: 114 (1) the studies were conducted in the field, including control and altered precipitation 115 (IPPT or DPPT); cropland and laboratory studies were excluded, 116 (2) the numbers of replicates (≥ 3), means, and standard deviations (SD) or standard errors 117 (SE) of C:N:P ratios were presented in the original paper, 118 (3) magnitude and duration of precipitation change were clearly described, 119 (4) the control and altered precipitation treatments were at the same initial experimental 120 conditions, 121 (5) for multifactorial experiments, other manipulation factors (e.g., warming, nutrient 122 addition, etc.) and their interactions were excluded in the precipitation treatments. 5 123 Using these criteria, a total of 386 IPPT and 442 DPPT independent paired sites 124 (Supporting Information Dataset S1) from 236 articles were selected for further analyses. For 125 the studies reported only the response of the element (C, N or P) concentration, we calculated 126 the corresponding element ratios. For the studies that do not provide standard deviation (SD), 127 we used the “Bracken1992” method to estimate it using “metagear” package in R software 128 (Benítez-López et al., 2017). The observations were collected mainly in North America, Europe 129 and Asia (Figure 1). The studies that fitted our criteria were located between 45.7° S and 76.5° 130 N, with MAT ranging from −11.3 to 28 ℃ and MAP from 81 to 3, 990 mm (Supporting 131 Information Dataset S1). The magnitude of precipitation treatments ranged from −100% to 132 +100%, the duration of precipitation treatments ranged from 1 to 15 years. 133 To standardize the level of precipitation manipulation in each study, all the operate levels 134 were converted to percentage of annual precipitation changes [ΔPPT (%)]. For IPPT, treatments 135 manipulations were implemented through two methods: 136 1) A constant water amount was added each year. For this situation, we recorded the annual 137 precipitation data of each research year, and then calculated ΔPPT. 138 2) The precipitation was collected from the water shelter, and then was discharged into the 139 IPPT experimental plots, or the water was added to a fixed proportion after each rainfall event. 140 In these cases, the ΔPPT were calculated based on Equation (1), of which PPT was the 141 precipitation during the manipulation period, the ΔPPT is the annual precipitation change. For 142 the DPPT that were conducted by the establishment of a rainout shelter, the ΔPPT was 143 calculated as the percentage of the rainout-shelter area to the total plot (PERshelter) if the DPPT 144 was performed throughout the year; or if the DPPT was seasonal, DPPT was calculated using 145 the method of (Hoover et al., 2018). In Equation (1), APPT is the annual amount of precipitation, 146 PPTTP was the total amount of precipitation among the experimental period. PPTTP 147 ΔPPT = PER shield × (1) APPT 6 148 Figure 1 149 150 2.2 |Factors affecting the response of C:N:P stoichiometry 151 Several factors affect the response of C:N:P stoichiometry of plants, soil and microbial 152 biomass to altered precipitation. These factors include climate variables (i.e., MAT, MAP), 153 ecosystem type (Ecosystem Type), change of precipitation levels (Magnitude), and duration of 154 precipitation manipulation (Duration). These variables were recorded directly from the original 155 or referenced papers. If the original climate information (MAT, MAP and potential 156 evapotranspiration) was not available, the global climate database (WorldClim, version 2.0, 157 http//:www.worldclim.org) was used. The humidity index (HI) was calculated as the ratio of 158 mean annual precipitation to mean annual potential evapotranspiration. The graphical data were 159 obtained digitalized by the free software Webplotdigitizer software (Burda et al., 2017) to 160 extract the means and SDs or SEs. 161 162 2.3 | Statistical analysis 163 We evaluate the responses ratio (RR) of C:N:P stoichiometry in each study followed the 164 methods of (Hedges et al., 1999) to altered precipitation: 165 RR = ln (Xe / Xc) = lnXe – lnXc (2) 166 The variance (v) of each RR was calculated as follows: 167 vi = (Se / Xe)2/ ne + (Sc / Xc)2 / nc (3) 168 where Xc and Xt are the means of a specific variable in the control and the treatment groups, 169 respectively; nc and nt are the number of replicates of the in the control and altered precipitation 170 groups, respectively. Se and Sc are the SD in the control and altered precipitation groups of each 171 observation, respectively. 172 The “metafor” package in R software was used to calculate the weighted response ratio (ln 7 173 RR++) and its 95% confidence interval (CI) using weighted regressions and random-effects 174 models. The weighted mean response ratio (RR++) was calculated using Equation (4): 175 RR++ = ∑𝑚 𝑘 𝑚 𝑘 𝑖=1 ∑𝑗=1 wij RRij /∑𝑖=1 ∑𝑗=1 wij (4) 176 where, m is the number of groups, k is the number of comparisons in the ith group, and wij 177 is the weight of each response ratio. The weight (wij) was calculated according to Equation (5): 178 wij= 1/(vi+τ2) (5) 179 Two variances are considered to calculate the weighted factors (wij) and RR in the mixed 180 effect model, including the within-study (v) and among-study (τ2) variances (Benítez-López et 181 al., 2017). 182 A variance linked to RR++ was calculated using Equation (6): 1 183 SE (RR++) =√∑m k (6) i=1 ∑j=1 wij 184 The weighted response ratio (RR++) and 95% confident interval (CI) were calculated by 185 “rma.mv” function in the “metafor” package. If 95% CI of weighted lnRR did not include zero, 186 the effect of precipitation manipulation was considered as significant (p <0.05). 187 The effect sizes of the terrestrial C:N:P stoichiometry were grouped by the type of 188 moderator variables (for example, the ecosystem type, humidity, and precipitation magnitude, 189 etc.) in response to the direction and the size of altered precipitation. To test whether the 190 response ratio between the sub-groups differ, we conducted the heterogeneity test (Qm test). 191 When the Qm value was significant (p <0.05), the response ratio between each group (ecosystem 192 type, humidity, and precipitation magnitude, etc.) was accepted as different. The weighted 193 response ratio and its 95% confidence interval (CI) were converted to the percentage of changes 194 by altered precipitation to facilitate explanation (as shown in the equation 7) and understanding 195 effects. The weighted RR++ greater than 0 indicates the positive effect of altered precipitation 196 on the analyzed variable, and lower than 0 means negative effects. 8 197 Percentage change = (eRR++ – 1) × 100% (7) 198 The sensitivity of each variable is standardized by the magnitude of precipitation change, 199 which helps to evaluate the response of plants, soil and microorganisms. The sensitivity was 200 calculated as follows: 𝑅𝑅 201 Sensitivity = (8) ΔPPT 202 The absolute value of sensitivity reflects the size of sensitivity, positive and negative 203 symbol represents the response direction. The variance (v, sensitivity) of sensitivity is 204 calculated using Equation (9). V𝑅𝑅 205 Vsensitivity = (9) ΔPPT2 206 The sensitive weighted average (sensitivity++) and the calculation method of its SE are the 207 same as RR++. 208 The single meta-regression was used to check the relationships between the RR and 209 continuous variables (e.g., magnitude and duration). To examine the heterogeneity of each 210 variable across subgroup categories (plants, soil, microorganisms), we calculated the between- 211 group heterogeneity (Qm) among ecosystem types (grassland, shrubland and forest), climate 212 humidity (dry and humid) and magnitude of precipitation (slight, medium and extreme). The 213 results of between-group heterogeneity of humidity were classified by HI = 0.65 (Table S2-S4). 214 The relative importance of all possible combinations of the environmental variables was 215 identified to analyze the most important predictors of the effects of altered precipitation on 216 C:N:P stoichiometry using “glmulti” package in R software (4.0). The relative importance of 217 each predictor was calculated the sum of Akaike weights containing all models including this 218 predictor using corrected Akaike's Information Criteria (AIC). These values can be considered 219 to the overall effect of each variable in all models. A cutoff of 0.8 was set to distinguish between 9 220 nonessential and important predictors. The funnel plots were used to explore the possibility of 221 publication bias (Figure S7-S12). 222 223 3 | Results 224 3.1 | Responses of C:N:P stoichiometry to altered precipitation 225 IPPT had no effects on leaf C, N and P concentrations, but elevated leaf and shoot C:P ratios 226 across all ecosystem types (Figure 2e). DPPT decreased leaf C:N ratio, but increased leaf N:P, 227 litter N (7.5%), root N (4%) and root N:P ratio across all ecosystem types (Figure 2b and f). 228 Compared with ambient precipitation, IPPT increased soil C (4.4%) and N (3.5%) concentration; 229 and C:P ratio, but DPPT reduced soil C (4.4%), N (5.2%) concentrations, and C:P ratio (Figure 230 2a, b, and e). IPPT increased microbial biomass C, N, and P by 13%, 15%, and 26%, 231 respectively (Figure 2a-c), but DPPT reduced microbial biomass N (12%) and N:P ratio (Figure 232 2b and f). 233 Figure 2 234 235 3.2 | Response of plant, soil and microbial C:N:P stoichiometry to altered precipitation in 236 dry and humid climates 237 IPPT and DPPT have opposite impacts on plant-soil-microbial C:N:P stoichiometry in dry and 238 humid areas (Figure 3). In dryland, IPPT increased leaf and soil C:P by 7.8% and 4.8% (Figure 239 3e), respectively. Similarly, IPPT increased C and N concentration of soil (4.6% and 3.7%) and 240 microbes (16% and 19%) in dryland. IPPT raised the C and N concentration in microbial 241 biomass 4-6 times greater than in soil in drylands (Figure 3a and b). DPPT had no effect on 242 plant, soil and microbial C:N:P except increased leaf N concentration in dryland (Figure 3b). 243 In humid climate, DPPT decreased soil N and N:P by 6.7%, 13%, respectively, but increased 244 leaf N and N:P by 19%, 13%, respectively (Figure 3b and f). 10 245 Figure 3 246 247 3.3 | Sensitivity of plant, soil and microbial C:N:P stoichiometry to altered precipitation 248 Leaf C:N and N:P were sensitive to medium DPPT because of N concentration was sensitive to 249 medium DPPT (Figure 4b, d and f), but leaf C:P was sensitive to medium IPPT (Figure 4e). 250 Soil C:P were sensitive to medium IPPT owing to the sensitivity of soil C to medium IPPT 251 (Figure 4a and e). Soil C, N, C:N, C:P and N:P were sensitive to medium DPPT (Figure 4a, b, 252 d-f), whereas soil N, C:P and N:P sensitive to slight DPPT (Figure 4a, e and f). In contrast, 253 microbial biomass C, N, and C:N were sensitive to slight IPPT (Figure 4a-d). Microbial biomass 254 C:N were sensitive to extreme DPPT the sensitivity of microbial biomass N (Figure 4b and d). 255 When the data is grouped by dry and humid ecosystems, the plants, soil, and 256 microorganisms respond differently to altered precipitation (Table S1). C and C:P of plants in 257 dry ecosystems were sensitive to medium IPPT, but not sensitive to DPPT (Table S1). In 258 contrast, N, C:N and N:P of plants in humid ecosystem were sensitive to medium DPPT, but 259 not sensitive to IPPT (Table S1). Soil C:N:P were majorly sensitive to medium IPPT in dry and 260 humid ecosystem. N, C:N and N:P of soil microorganisms were sensitive to slight IPPT in dry 261 and humid ecosystem, but not sensitive to DPPT (Table S1). In addition, microbes were majorly 262 sensitive to extreme IPPT in dry ecosystem (Table S1). 263 Figure 4 264 265 3.4 | Factors influencing the response of plant-soil-microbial C:N:P stoichiometry to 266 altered precipitation 267 Model (summed Akaike weights) selection analysis demonstrated that plant N, C:N and N:P 268 were best explained by MAP and MAT, but none of the variables reached the Akaike threshold 269 of 0.8 on C, P and C:P to altered precipitation (Table 1). In contrast, precipitation magnitude 11 270 was the most important factor affecting soil and microbial C:N, C:P and N:P ratios. 271 Linear regression analysis confirmed that leaf and root N concentrations as well as N:P 272 ratio elevated with increasing MAP, but leaf and root C:N declined with increasing MAP 273 (Figure S1b, c and f). Soil C, N, C:P and microbial biomass C and N:P elevated with the 274 increasing precipitation magnitude, but microbial biomass C:N decreased with increasing 275 precipitation magnitude (Figure S2a,b, d-f). 276 277 4 | Discussion 278 4.1 | Response of C:N:P stoichiometry to altered precipitation 279 Altered precipitation affects plant and microbial growth by changing the soil water availability, 280 and altering the fate of C sequestration (Abbasi et al., 2020b; Beier et al., 2012; Rousk et al., 281 2013). DPPT decreased leaf C:N, and increased leaf N:P, litter N, root N and N:P. DPPT 282 decreased soil water content, and consequently the capacity of uptake of P more than the 283 capacity to take up N by plants, thus diminishing proportionally more the leaching losses of soil 284 N than of P and resulting in increased N:P ratio in plant communities. Higher plant N:P ratio 285 can favor species with low growth rate, and is favorable for adapting to arid environments 286 (Sardans et al., 2021). Decreased C:N ratio in leaves, litters and roots by DPPT were mostly 287 attributed to a loss of N use efficiency due to more severe water limitation and decreased 288 activity of key enzymes involved in photosynthesis (e.g., ribulose 1,5-bisphosphate 289 carboxylase/oxygenase) in plants (Flexas et al., 2006; Zhou et al., 2016). DPPT decreased the 290 leaf C:N ratio in forest ecosystems and humid regions (Figure 3d and 6d), demonstrating that 291 N is accumulated in leaves of trees (Figure 3). The possible mechanisms were that plants have 292 closer symbiotic associations with AM fungi to increase N and P uptake and water use 293 efficiency under drought stress (Figure 6; Mariotte et al., 2017). This situation reduced soil 294 nutrient concentrations, which was supported by that DPPT decreased soil C, N and C:P 12 295 compared with ambient precipitation owing to P uptake is more seriously limited than that of 296 N under drought (Figure 2a, b and e). Similarly, DPPT reduced microbial biomass N and N:P 297 (Figure 2b and f), and the possible mechanism is that drought led to the dieback of microbial 298 cells, and decreased microbial biomass N concentration (Gu et al., 2019). These changes 299 prompted the decline of microbial N use efficiency, and decreased microbial activities (Zeglin 300 et al., 2013). 301 Compared with ambient precipitation, IPPT effects on plant, soil and microbial C:N:P 302 stoichiometry were of minor importance than the observed under DPPT except elevated leaf 303 and shoot C:P (Figure 2 a-f). This could be mainly related to that IPPT increased P use efficiency, 304 litter inputs and P mineralization consistent with previous studies (Austin et al., 2000), which 305 altered the balance of C and P in leaf and shoot. IPPT increased soil and microbial biomass C 306 and N concentration (Figure 2a and b) because IPPT increased plant biomass (Figure S4), and 307 provided more litter input belowground and rhizodeposition (Figure 6, Clark et al., 2009; 308 Khalili et al., 2016). The increased C resource enhanced the availability of soluble substrates 309 for microbial communities (Figure S5, Clark et al., 2009; Khalili et al., 2016). Subgroup 310 analysis found that IPPT mainly affected microbial biomass C:N:P in dry areas and grasslands 311 (Figure 3), indicating microbial communities in dry ecosystems are susceptive to increased 312 precipitation. Increased precipitation increased litter input and decomposition rates, which 313 provided the nearest photosynthate to soil microbes (Austin & Vitousek, 2000; Hao et al., 2017). 314 Medium precipitation stimulates exudation of root metabolites, which act as enzymes leading 315 to rapid decomposition of organic material and release of labile C (Clark et al., 2009; Zhou et 316 al., 2018). IPPT declined the microbial biomass C:N ratio, which was related to the shifts in 317 microbial communities to higher precipitation (Zhou et al., 2018). Increasing soil water 318 availability increased the growth of bacteria with high C:N (Hess & Austin, 2017; Li et al., 319 2020). 13 320 Figure 6 321 Figure 7 322 Plants and microbes can maintain or alter their biomass nutrient composition to adapt to 323 changes of nutrient availability in soil by precipitation changes. Leaf N:P increased with soil 324 N:P, indicating that the leaf N:P is plastic (S = 0.78, Figure 7f). Leaves with high N and P 325 concentrations tend to be short lived and structurally fragile, with high specific leaf area and 326 photosynthetic activity (Sardans et al., 2012). Strong stoichiometric homeostasis of soil 327 microbes under altered precipitation with the non-significant relationships between soil C:N:P 328 stoichiometry and microbial biomass C:N:P (Figure 7d-f). Soil microbes actively adjusted 329 physiologically to adapt to changes in soil N and P resources despite changing soil moisture 330 (Yang et al., 2011; Xu et al., 2013). Homeostasis regulation is essential for microbial growth, 331 which is a critical physiological regulation in maintaining microbial biomass production (Heuck 332 et al., 2015; Sterner & Elser, 2002). Most soil microbes are heterotrophic and thought to exhibit 333 stronger homeostasis behavior at the organic level than autotrophs (Clark et al., 2009). 334 Decreased precipitation increases fungi and decrease bacterial abundance, because fungi have 335 stronger anti-drought stress cell walls and extended hyphal structure to efficiently extract water, 336 leading to higher C:N (Ullah et al., 2021). Therefore, altered precipitation shift the microbial 337 homeostasis through altering microbial community composition (Clark et al., 2009; Delgado‐ 338 Baquerizo et al., 2017). 339 340 4.2 | Sensitivities of C:N:P stoichiometry to precipitation changes 341 Leaf N concentration, C:N and N:P ratios were sensitive to medium DPPT (Figures 4b, d and 342 f), further confirmed that DPPT mainly affected the N cycle of plants. Drought can increase 343 plant N uptake in humid areas (Wu et al., 2022), thus with high sensitivity of leaf N 344 concentration, C:N and N:P ratios to DPPT in humid regions (Table S1). One important possible 14 345 mechanism is that drought can diminish N losses by leaching (Austin & Vitousek, 2000; Hao 346 et al., 2017). Specially, humid area has high soil moisture, and medium drought could increase 347 soil aeration can favor root activity and increase N uptake by roots from soil (Deng et al., 2021; 348 Farooq et al., 2009; Hertel et al., 2013). Subgroup analysis further revealed that soil C and N 349 concentrations, and C:N, C:P and N:P ratios were sensitive to medium DPPT (Table S1). These 350 results were consistent with the response of forests to DPPT (Figure S6), suggesting that forests 351 are more vulnerable to drought. The possible mechanism is that long-term drought increased 352 the mortality of single trees, and forests need more water for their greater stand biomass are the 353 biomes more affected by drought, thus it takes longer to fully restore the level before the stress 354 owing to the stomatal and/or photochemical limitation. These results were consistent with the 355 decreased annual precipitation in most subtropical regions (Harper et al., 2005). 356 Soil microbial biomass C and N concentrations, and C:N ratio were sensitive to slight IPPT 357 (Figure 4a-d), and microbial biomass N concentration and C:N ratio were also sensitive to 358 extreme DPPT (Figure 4b and d). Similarly, microbial biomass C, N and P concentrations and 359 C:N ratio respond to IPPT, not to DPPT (Figure 5), and the response of microbial biomass N 360 concentration to IPPT (15%) is greater than that to DPPT (12%). These results were consistent 361 with the prediction of the double asymmetric model proposed by Knapp et al. (2017). Soil water 362 availability after severe drought reduce solute diffusivity and inhibit microbial growth due to 363 the inability to access substrates under dry conditions (Ren et al., 2017; Schimel et al., 2007). 364 In dry areas, DPPT reduced soil microbial activities, which in turn are more susceptible to water 365 availability (Schimel et al., 2007; She et al., 2018). 366 Soil microbes (N, C:N and N:P) were sensitive to slight IPPT in dry and humid ecosystem 367 (Table S1). Slight IPPT can increase water availability and relieve water stresses for soil 368 microorganisms, promoting microbial nutrient absorption and growth (Huxman et al., 2004). 369 Small precipitation events, especially in dry ecosystems, eliminated water-induced substrate 15 370 limitation and increases microbial biomass C, N and P concentrations as observed in previous 371 studies (Davidson et al., 2012; Evans & Wallenstein, 2014; Wang et al., 2021). Therefore, 372 precipitation change regimes alter water sensitivity and ultimately influencing microbial growth 373 in terrestrial ecosystems. 374 Figure 5 375 376 4.3 | Driving factors of C:N:P stoichiometry response to altered precipitation 377 Plant, soil and microbial C:N:P stoichiometry responses to altered precipitation were affected 378 by climate factors and precipitation magnitude. This study found that the response ratio of leaf 379 and root N concentration and C:N ratio were related to MAP (Figure S1b and d), indicating that 380 plant N was closely related to MAP changes to altered precipitation. This relationship can be 381 attributed to the gradients of ecosystem N cycle along environmental and experimental factors 382 (Xu et al., 2020; Yue et al., 2017). Plant growth depends on the net photosynthetic rate, 383 photosynthetic efficiency, which mainly directly affected by climate factors (Farooq etal., 2009). 384 N is a component of chlorophyll and an essential element for photosynthetic enzymes, and MAP 385 can directly or indirectly increase N absorbed by the roots and their ability to transport to plant 386 organs by changing vegetation composition and soil biogeographic processes (Sardans et al., 387 2012; Yue et al., 2019). In addition, the N cycle between plants and soil is faster in low-latitude 388 bio communities with high MAP, accompanied by high N gas or leaching losses, and biological 389 nitrogen fixation capacity, such as subtropical and tropical forests (Deng et al, 2018). Different 390 ecosystems dominate species distributions due to differences in climate factors. Ecosystem type 391 influenced plant N concentration (Table S2, Qm = 9.67, p = 0.02) and C:N ratio (Table S2, Qm 392 = 11.87, p = 0.008) ratios under DPPT, but not in grasslands and deserts ecosystems (Figure 393 S6). This differential response is that plant functional types (e.g., nitrogen-fixing species) differ 394 between ecosystem types (Pugnaire et al., 2019). Consequently, leaf stoichiometry to altered 16 395 precipitation is specific across sites with climate conditions. 396 Leaf N, soil C and microbial biomass C concentrations, C:N, and N:P ratios were related 397 to the magnitude of altered precipitation (Figure S2a, d and f). These results were consistent 398 with previous meta-analysis (Knapp et al., 2015; Zhou et al., 2018). Increased precipitation 399 elevated soil water availability, and accelerates microbial decomposition of soil organic matter 400 in arid ecosystems (Wang et al., 2021; Zeglin et al., 2013). Decreased precipitation leads to a 401 linear relationship between plant and microbial responses with precipitation changes (Abbasi 402 et al., 2020; Dijkstra et al., 2012). These relationships result from the normal range of treatment 403 magnitudes in most precipitation experiments worldwide (Wang et al., 2021). When combined 404 with extreme precipitation, the responses of other plant-microbial processes deviated from the 405 linear relationship with precipitation variation amplitude (Luo et al., 2017). Extreme drought 406 results in impaired photosynthesis, decreased microbial nutrient mineralization, and even death 407 of plants and microorganisms, while extreme precipitation brings excess water, restricted soil 408 oxygen, and limit plant and soil microbial growth (Wang et al., 2020). Therefore, more extreme 409 precipitation (≥67%) and multiple levels of precipitation experiment should be performed to 410 assess whether these linear relationships hold under extreme conditions in the future. 411 412 4.4 | Implications of precipitation changes for C cycle and future research 413 The N-related variables of plants, soil and microbes were sensitive to altered precipitation 414 (Figure 6), which suggested that N cycles altered the growth and metabolism of plants and 415 microbes. Previous studies have demonstrated that N limited plant primary productivity and C 416 sequestration in various ecosystems (Elser et al., 2007; Vitousek et al., 2010). In general, N 417 rapidly accumulates from the atmosphere through biological N fixation that dominate the early 418 stages of ecosystem development (Delgado‐Baquerizo et al., 2017; Sardans et al., 2021). But, 419 N accumulates slowly from the atmosphere through atmospheric deposition and biological 17 420 fixation and tend to become less limiting in more advanced successional stages (DeMott & 421 Pape, 2005; Luo et al., 2017). Most free-living microbial communities are N-limited, and 422 secreted N acquiring-extracellular enzymes that catalyzed organic matter decomposition 423 (Hewins et al., 2016; Mariotte et al., 2017). These processes can construct plant and microbial 424 cell walls, depolymerize macromolecules, and ultimately provide soluble substrates for 425 microbial assimilation (Henry et al., 2005; Kaiser et al., 2014). Therefore, altered precipitation 426 not only affect the N-related leaf photosynthesis, and net primary productivity (Nielsen & Ball, 427 2015), but also influence the soil extracellular enzyme activities by microbes involved in N 428 mineralization, and ultimately affect the accumulation of soil organic carbon (Waldrop et al., 429 2004; Yuan & Chen, 2015). 430 Precipitation manipulation experiments have been conducted across the world, but there 431 are some limitations and uncertainties in experimental studies of precipitation change. Available 432 data for some parameters or processes are still limited, for example, the number of studies of 433 leaf C, N, and P concentrations, C:N, C:P, and N:Pratios in shrubland ecosystems is less than 434 10 (Figure 3). The distribution of sampling points is also an important factor that increases the 435 uncertainty of meta-analysis. Most of the experiments are in the northern Hemisphere, such as 436 China, Europe and the United States (Figure 1). Savanna ecosystems in the Southern 437 Hemisphere, particularly from tropical and subtropical regions of Africa, contributed to 438 uncertainty in the analysis of this study. Overall, although the meta-analysis identified the 439 importance of ecosystem type, rainfall intensity, and climate on plant-soil-microbial C:N:P 440 stoichiometry, most of the studies had short experimental periods (< 10 years). As a result, a 441 comprehensive understanding of how precipitation affects plant-soil-microbial nutrient 442 dynamics over time scales of more than a decade is still lacking (Yue et al., 2019). The rewetting 443 process during the precipitation treatment period (early or late vegetation season), precipitation 444 period and the first precipitation after precipitation also had important effects on plant-soil- 18 445 microbial C:N:P stoichiometry dynamics. 446 447 5 | Conclusions 448 The systematic response of C, N and P concentration and consequently the C:N:P stoichiometry 449 to altered precipitation across the globe provides unparalleled insights into ecosystem 450 biogeochemical processes by increased and decreased precipitation. Drought increased leaf C:N 451 ratio, but increased leaf N:P ratio on a global scale. Increased precipitation elevated soil C, and 452 N concentrations, while DPPT decreased them. Increased precipitation increased microbial 453 biomass C, N and P concentrations, whereas drought reduced microbial biomass N 454 concentration and N:P ratio due to the changes in microbial community composition. Plant 455 communities were more sensitive to decreased precipitation than to increased precipitation, 456 especially in humid areas. Soil and microbes were more sensitive to increased precipitation than 457 to decreased precipitation, especially in dry areas. MAP and precipitation magnitude regulated 458 the response of plant-soil-microbial C:N:P stoichiometry to altered precipitation. Consequently, 459 incorporating the specific aspects of precipitation intensity and N-related traits into the 460 framework of global climate predictions could tribute to a deeper understanding of precipitation 461 mediated C dynamics. 462 Acknowledgements 463 This study was supported by the Second Tibetan Plateau Scientific Expedition and Research 464 Program (2019QZKK0603), the National Natural Science Foundation of China (U2243225, 465 42277471), the Strategic Priority Research Program of the Chinese Academy of Sciences 466 (XDA23070201), and the Scientific and Technological Innovation Project of Shaanxi Forestry 467 Academy of Sciences (SXLK2021-0206), the Key Research and Development Program of 468 Shaanxi Province (2021ZDLSF05-02), and the RUDN University Strategic Academic 469 Leadership Program. 19 470 471 Author contributions 472 J.L. and L.D. designed the research, J.L. and L.D. collected the data, J.L. performed the research, 473 J.L., L.D., J.P., and Y.K. analysed the data, J.L., L.D. and J.W. visualized the figures, and J.W., 474 L.D. J.P., J. S., S.Z., and Y.K. wrote the paper. 475 476 Declaration of Competing Interest 477 All the authors declare no competing interests. 478 Supplementary information 479 Appendix table S1-S4 and Appendix figures S1-S12. 480 481 References 482 Abbasi, A. O., Salazar, A., Oh, Y., Reinsch, S., del Rosario Uribe, M., Li, J., Rashid, I., & Dukes, J. S. (2020). 483 Reviews and syntheses: Soil responses to manipulated precipitation changes–an assessment of meta- 484 analyses. Biogeosciences, 17, 3859–3873. 485 Austin, A. T., & Vitousek, P. M. (2000). Precipitation, decomposition and litter decomposability of 486 Metrosideros polymorpha in native forests on Hawai’i. Journal of Ecology, 88, 129–138. 487 Beier, C., Beierkuhnlein, C., Wohlgemuth, T., Penuelas, J., Emmett, B., Körner, C., de Boeck, H., Christensen, 488 J. H., Leuzinger, S., & Janssens, I. A. (2012). Precipitation manipulation experiments–challenges 489 and recommendations for the future. Ecology Letters, 15, 899–911. 490 Benítez-López, A., Alkemade, R., Schipper, A., Ingram, D., Verweij, P., Eikelboom, J., & Huijbregts, M. 491 (2017). The impact of hunting on tropical mammal and bird populations. Science, 356, 180–183. 492 Burda, B. U., O’Connor, E. A., Webber, E. M., Redmond, N., & Perdue, L. A. (2017). Estimating data from 493 figures with a Web‐based program: Considerations for a systematic review. Research Synthesis 494 Methods, 8, 258–262. 495 Canarini, A., Kiær, L. P., & Dijkstra, F. A. (2017). Soil carbon loss regulated by drought intensity and 496 available substrate: A meta-analysis. Soil Biology and Biochemistry, 112, 90–99. 497 Clark, J. S., Campbell, J. H., Grizzle, H., Acosta-Martìnez, V., & Zak, J. C. (2009). Soil microbial community 498 response to drought and precipitation variability in the Chihuahuan Desert. Microbial Ecology, 57, 499 248–260. 500 Davidson, E. A., Samanta, S., Caramori, S. S., & Savage, K. (2012). The D ual A rrhenius and M ichaelis–M 501 enten kinetics model for decomposition of soil organic matter at hourly to seasonal time scales. 502 Global Change Biology, 18, 371–384. 503 Delgado‐Baquerizo, M., Reich, P. B., Khachane, A. N., Campbell, C. D., Thomas, N., Freitag, T. E., Abu Al‐ 504 Soud, W., Sørensen, S., Bardgett, R. D., & Singh, B. K. (2017). It is elemental: soil nutrient 505 stoichiometry drives bacterial diversity. Environmental Microbiology, 19, 1176–1188. 506 DeMott, W. R., & Pape, B. J. J. O. (2005). Stoichiometry in an ecological context: testing for links between 507 Daphnia P-content, growth rate and habitat preference. 142, 20–27. 508 Deng, L., Peng, C., Kim, D. G., Li, J., Liu, Y., Hai, X., Liu, Q., Huang, C., Shangguan, Z., & Kuzyakov, Y. 509 (2021). Drought effects on soil carbon and nitrogen dynamics in global natural ecosystems. Earth- 510 Science Reviews, 214, 103501. 511 Deng, M., Liu, L., Jiang, L., Liu, W., Wang, X., Li, S., Yang, S., & Wang, B. (2018). Ecosystem scale trade- 512 off in nitrogen acquisition pathways. Nature Ecology & Evolution, 2(11), 1724–1734. 513 Dijkstra, F. A., Pendall, E., Morgan, J. A., Blumenthal, D. M., Carrillo, Y., LeCain, D. R., Follett, R. F., & 514 Williams, D. G. (2012). Climate change alters stoichiometry of phosphorus and nitrogen in a 515 semiarid grassland. New Phytologist, 196, 807–815. 516 Elser, J. J., Bracken, M. E., Cleland, E. E., Gruner, D. S., Harpole, W. S., Hillebrand, H., Ngai, J. T., Seabloom, 20 517 E. W., Shurin, J. B., & Smith, J. E. (2007). Global analysis of nitrogen and phosphorus limitation of 518 primary producers in freshwater, marine and terrestrial ecosystems. Ecology Letters, 10, 1135–1142. 519 Evans, S. E., & Wallenstein, M. D. (2014). Climate change alters ecological strategies of soil bacteria. 520 Ecology Letters, 17, 155–164. 521 Farooq, M., Wahid, A., Kobayashi, N., Fujita, D., & Basra, S. (2009). A. 2009. Plant drought stress: Effects, 522 mechanisms and management. Agronomy for Sustainable Development, 29, 185–212. 523 Flexas, J., Ribas‐Carbó, M., Bota, J., Galmés, J., Henkle, M., Martínez‐Cañellas, S., & Medrano, H. 524 (2006). Decreased Rubisco activity during water stress is not induced by decreased relative water 525 content but related to conditions of low stomatal conductance and chloroplast CO2 concentration. 526 New Phytologist, 172, 73–82. 527 Gu, Y., Wei, Y., Xiang, Q., Zhao, K., Yu, X., Zhang, X., Li, C., Chen, Q., Xiao, H., & Zhang, X. (2019). C: 528 N ratio shaped both taxonomic and functional structure of microbial communities in livestock and 529 poultry breeding wastewater treatment reactor. Science of the Total Environment, 651, 625–633. 530 Hao, Y., Zhou, C., Liu, W., Li, L., Kang, X., Jiang, L., Cui, X., Wang, Y., Zhou, X., & Xu, C. (2017). 531 Aboveground net primary productivity and carbon balance remain stable under extreme precipitation 532 events in a semiarid steppe ecosystem. Agricultural and Forest Meteorology, 240, 1–9. 533 Harper, C. W., Blair, J. M., Fay, P. A., Knapp, A. K., & Carlisle, J. D. (2005). Increased rainfall variability 534 and reduced rainfall amount decreases soil CO2 flux in a grassland ecosystem. Global Change 535 Biology, 11, 322–334. 536 He, M., & Dijkstra, F. A. (2014). Drought effect on plant nitrogen and phosphorus: a meta‐analysis. New 537 Phytologist, 204, 924–931. 538 Hedges, L. V., Gurevitch, J., & Curtis, P. S. (1999). The meta‐analysis of response ratios in experimental 539 ecology. Ecology, 80, 1150–1156. 540 Henry, H. A. L., Cleland, E. E., Field, C. B., & Vitousek, P. M. (2005). Interactive effects of elevated CO2, N 541 deposition and climate change on extracellular enzyme activity and soil density fractionation in a 542 California annual grassland. Oecologia, 142, 465–473. 543 Hertel, D., Strecker, T., Müller‐Haubold, H., & Leuschner, C. (2013). Fine root biomass and dynamics in 544 beech forests across a precipitation gradient–is optimal resource partitioning theory applicable to 545 water‐limited mature trees? Journal of Ecology, 101, 1183–1200. 546 Hess, L. J. T., & Austin, A. T. (2017). Pine afforestation alters rhizosphere effects and soil nutrient turnover 547 across a precipitation gradient in Patagonia, Argentina. Plant and Soil, 415, 449-464. 548 Heuck, C., Weig, A., & Spohn, M. (2015). Soil microbial biomass C:N:P stoichiometry and microbial use of 549 organic phosphorus. Soil Biology and Biochemistry, 85, 119–129. 550 Hewins, D. B., Broadbent, T., Carlyle, C. N., & Bork, E. W. (2016). Extracellular enzyme activity response 551 to defoliation and water addition in two ecosites of the mixed grass prairie. Agriculture, Ecosystems 552 & Environment, 230, 79–86. 553 Hoover, D. L., Wilcox, K. R., & Young, K. E. (2018). Experimental droughts with rainout shelters: a 554 methodological review. Ecosphere, 9, e02088. 555 Huxman, T. E., Snyder, K. A., Tissue, D., Leffler, A. J., Ogle, K., Pockman, W. T., Sandquist, D. R., Potts, D. 556 L., & Schwinning, S. (2004). Precipitation pulses and carbon fluxes in semiarid and arid ecosystems. 557 Oecologia, 141, 254–268. 558 Kaiser, C., Franklin, O., Dieckmann, U., & Richter, A. (2014). Microbial community dynamics alleviate 559 stoichiometric constraints during litter decay. Ecology Letters, 17, 680–690. 560 Khalili, B., Ogunseitan, O. A., Goulden, M. L., & Allison, S. D. (2016). Interactive effects of precipitation 561 manipulation and nitrogen addition on soil properties in California grassland and shrubland. Applied 562 Soil Ecology, 107, 144–153. 563 Knapp, A. K., Ciais, P., & Smith, M. D. (2017). Reconciling inconsistencies in precipitation-productivity 564 relationships: implications for climate change. New Phytologist, 214, 41–47. 565 Knapp, A. K., Hoover, D. L., Wilcox, K. R., Avolio, M. L., Koerner, S. E., La Pierre, K. J., Loik, M. E., Luo, 566 Y., Sala, O. E., & Smith, M. D. (2015). Characterizing differences in precipitation regimes of 567 extreme wet and dry years: implications for climate change experiments. Global Change Biology, 568 21, 2624–2633. 569 Li, J., Shangguan, Z., & Deng, L. (2020). Dynamics of soil microbial metabolic activity during grassland 570 succession after farmland abandonment. Geoderma, 363, 114167. 571 Luo, Y., Jiang, L., Niu, S., & Zhou, X. (2017). Nonlinear responses of land ecosystems to variation in 572 precipitation. New Phytologist, 214, 5–7. 21 573 Ma, H. K., Bai, G. Y., Sun, Y., Kostenko, O., Zhu, X., Lin, S., Ruan, W. B., Zhao, N. X., & Bezemer, T. M. 574 (2016). Opposing effects of nitrogen and water addition on soil bacterial and fungal communities in 575 the Inner Mongolia steppe: A field experiment. Applied Soil Ecology, 108, 128–135. 576 Mariotte, P., Canarini, A., & Dijkstra, F. A. (2017). Stoichiometric N: P flexibility and mycorrhizal symbiosis 577 favour plant resistance against drought. Journal of Ecology, 105, 958–967. 578 Maurer, G. E., Hallmark, A. J., Brown, R. F., Sala, O. E., & Collins, S. L. (2020). Sensitivity of primary 579 production to precipitation across the United States. Ecology Letters, 23, 527–536. 580 McGroddy, M. E., Daufresne, T., & Hedin, L. O. (2004). Scaling of C: N: P stoichiometry in forests 581 worldwide: Implications of terrestrial redfield‐type ratios. Ecology, 85, 2390–2401. 582 Nielsen, U. N., & Ball, B. A. (2015). Impacts of altered precipitation regimes on soil communities and 583 biogeochemistry in arid and semi‐arid ecosystems. Global Change Biology, 21, 1407–1421. 584 Prommer, J., Walker, T. W., Wanek, W., Braun, J., Zezula, D., Hu, Y., Hofhansl, F., & Richter, A. (2019). 585 Increased microbial growth, biomass, and turnover drive soil organic carbon accumulation at higher 586 plant diversity. Global Change Biology, 26, 669–681. 587 Pugnaire, F. I., Morillo, J. A., Peñuelas, J., Reich, P. B., Bardgett, R. D., Gaxiola, A., Wardle, D. A., & Van 588 Der Putten, W. H. (2019). Climate change effects on plant-soil feedbacks and consequences for 589 biodiversity and functioning of terrestrial ecosystems. Science Advances, 5, eaaz1834. 590 Ren, C., Zhao, F., Shi, Z., Chen, J., Han, X., Yang, G., Feng, Y., & Ren, G. (2017). Differential responses of 591 soil microbial biomass and carbon-degrading enzyme activities to altered precipitation. Soil Biology 592 and Biochemistry, 115, 1–10. 593 Rousk, J., Smith, A. R., & Jones, D. L. (2013). Investigating the long‐term legacy of drought and warming 594 on the soil microbial community across five E uropean shrubland ecosystems. Global Change 595 Biology, 19, 3872–3884. 596 Sardans, J., Janssens, I. A., Ciais, P., Obersteiner, M., & Peñuelas, J. (2021). Recent advances and future 597 research in ecological stoichiometry. Perspectives in Plant Ecology, Evolution and Systematics, 50, 598 125611. 599 Sardans, J., Rivas-Ubach, A., & Peñuelas, J. (2012). The C:N:P stoichiometry of organisms and ecosystems 600 in a changing world: a review and perspectives. Perspectives in Plant Ecology, Evolution and 601 Systematics, 14, 33–47. 602 Schimel, J., Balser, T. C., & Wallenstein, M. (2007). Microbial stress‐response physiology and its 603 implications for ecosystem function. Ecology, 88, 1386–1394. 604 She, W., Bai, Y., Zhang, Y., Qin, S., Feng, W., Sun, Y., Zheng, J., & Wu, B. (2018). Resource availability 605 drives responses of soil microbial communities to short-term precipitation and nitrogen addition in 606 a desert shrubland. Frontiers in Microbiology, 9, 186. 607 Sierra, C. A., Malghani, S., & Loescher, H. W. (2017). Interactions among temperature, moisture, and oxygen 608 concentrations in controlling decomposition rates in a boreal forest soil. Biogeosciences, 14, 703– 609 710. 610 Smith, C., Baker, J. C. A., & Spracklen, D. V. (2023). Tropical deforestation causes large reductions in 611 observed precipitation. Nature, 1–6. 612 Sterner, R. W., & Elser, J. J. (2002). Ecological stoichiometry: the biology of elements from molecules to the 613 biosphere: Princeton university press. 614 Stocker, T. (2014). Climate change 2013: the physical science basis: Working Group I contribution to the 615 Fifth assessment report of the Intergovernmental Panel on Climate Change: Cambridge university 616 press. 617 Thakur, M. P., Milcu, A., Manning, P., Niklaus, P. A., Roscher, C., Power, S., Reich, P. B., Scheu, S., Tilman, 618 D., & Ai, F. (2015). Plant diversity drives soil microbial biomass carbon in grasslands irrespective 619 of global environmental change factors. Global Change Biology, 21, 4076–4085. 620 Ullah, M. R., Carrillo, Y., Dijkstra, F. A. (2021). Drought-induced and seasonal variation in carbon use 621 efficiency is associated with fungi: bacteria ratio and enzyme production in a grassland ecosystem. 622 Soil Biology and Biochemistry, 155, 108159. 623 Vitousek, P. M., Porder, S., Houlton, B. Z., & Chadwick, O. A. (2010). Terrestrial phosphorus limitation: 624 mechanisms, implications, and nitrogen-phosphorus interactions. Ecological Applications, 20, 5–15. 625 Waldrop, M. P., Zak, D. R., Sinsabaugh, R. L., Gallo, M., & Lauber, C. (2004). Nitrogen deposition modifies 626 soil carbon storage through changes in microbial enzymatic activity. Ecological Applications, 14, 627 1172–1177. 628 Wang, J., Tian, D., Knapp, A. K., Chen, H. Y., Luo, Y., Li, Z., Hou, E., Huang, X., Jiang, L., & Niu, S. (2021). 22 629 Precipitation manipulation and terrestrial carbon cycling: The roles of treatment magnitude, 630 experimental duration and local climate. Global Ecology and Biogeography, 30, 1909–1921. 631 Wang, P., Huang, K., & Hu, S. (2020). Distinct fine‐root responses to precipitation changes in herbaceous 632 and woody plants: a meta‐analysis. New Phytologist, 225, 1491–1499. 633 Wilcox, K. R., Shi, Z., Gherardi, L. A., Lemoine, N. P., Koerner, S. E., Hoover, D. L., Bork, E., Byrne, K. 634 M., Cahill Jr, J., & Collins, S. L. (2017). Asymmetric responses of primary productivity to 635 precipitation extremes: a synthesis of grassland precipitation manipulation experiments. Global 636 Change Biology, 23, 4376–4385. 637 Wu, Q., Yue, K., Ma, Y., Heděnec, P., Cai, Y., Chen, J., Zhang, H., Shao, J., Chang, S. X., & Li, Y. (2022). 638 Contrasting effects of altered precipitation regimes on soil nitrogen cycling at the global scale. 639 Global Change Biology, 28, 6679–6695. 640 Wu, Z., Dijkstra, P., Koch, G. W., Peñuelas, J., & Hungate, B. A. (2011). Responses of terrestrial ecosystems 641 to temperature and precipitation change: A meta‐analysis of experimental manipulation. Global 642 Change Biology, 17, 927–942. 643 Xu, S., Geng, W., Sayer, E. J., Zhou, G., Zhou, P., & Liu, C. (2020). Soil microbial biomass and community 644 responses to experimental precipitation change: A meta-analysis. Soil Ecology Letters, 2, 93–103. 645 Yuan, Z., & Chen, H. Y. (2015). Decoupling of nitrogen and phosphorus in terrestrial plants associated with 646 global changes. Nature Climate Change, 5, 465–469. 647 Yue, K., Fornara, D. A., Yang, W., Peng, Y., Li, Z., Wu, F., & Peng, C. (2017). Effects of three global change 648 drivers on terrestrial C: N: P stoichiometry: a global synthesis. Global Change Biology, 23, 2450– 649 2463. 650 Yue, K., Peng, Y., Fornara, D. A., Van Meerbeek, K., Vesterdal, L., Yang, W., Peng, C., Tan, B., Zhou, W., & 651 Xu, Z. (2019). Responses of nitrogen concentrations and pools to multiple environmental change 652 drivers: A meta‐analysis across terrestrial ecosystems. Global Ecology and Biogeography, 28, 690– 653 724. 654 Zeglin, L., Bottomley, P. J., Jumpponen, A., Rice, C., Arango, M., Lindsley, A., McGowan, A., Mfombep, P., 655 & Myrold, D. (2013). Altered precipitation regime affects the function and composition of soil 656 microbial communities on multiple time scales. Ecology, 94, 2334–2345. 657 Zhou, X., Zhou, L., Nie, Y., Fu, Y., Du, Z., Shao, J., Zheng, Z., & Wang, X. (2016). Similar responses of soil 658 carbon storage to drought and irrigation in terrestrial ecosystems but with contrasting mechanisms: 659 A meta-analysis. Agriculture, Ecosystems & Environment, 228, 70–81. 660 Zhou, Z., Wang, C., & Luo, Y. (2018). Response of soil microbial communities to altered precipitation: A 661 global synthesis. Global Ecology and Biogeography, 27, 1121–1136. 662 23 663 Tables and Figures: 664 Table 1 Model-averaged importance of the predictors of the effects of altered precipitation on plant, soil, 665 and microbial C, N, P, concentrations, C:N, C:P and N:P ratios. Ecosystem Variables MAT MAP HI Duration Magnitude type C 0.2324 0.3457 0.2889 0.1026 0.2166 0.5436 N 0.9919 0.9996 0.9853 0.1090 0.3604 0.4409 P 0.2267 0.3862 0.535 0.086 0.2454 0.6695 Leaf C:N 0.4691 0.8228 0.6106 0.2598 0.5273 0.8051 C:P 0.3468 0.3479 0.2331 0.0880 0.2118 0.6027 N:P 0.8412 0.8123 0.4925 0.3584 0.6629 0.2339 C 0.6647 0.4293 0.3293 0.5690 0.2681 1 N 0.5464 0.3439 0.5094 0.9676 0.2612 0.9957 P 0.2475 0.5597 0.3501 0.6661 0.3780 0.9629 Soil C:N 0.2491 0.2894 0.4183 0.0747 0.4285 0.7165 C:P 0.2919 0.4678 0.5714 0.1984 0.3314 1 N:P 0.2959 0.4393 0.3565 0.1581 0.3421 1 C 0.4381 0.3453 0.3092 0.8624 0.2717 1 N 0.2836 0.2810 0.2668 0.2118 0.2822 1 Microbial P 0.2858 0.2737 0.2628 0.1295 0.3439 0.8116 biomass C:N 0.2810 0.4105 0.3440 0.2895 0.3242 1 C:P 0.0074 0.0002 0.1786 0.0542 0.0518 0.0006 N:P 0.2375 0.2376 0.2323 0.0650 0.4812 0.3898 666 Importance is estimated from the sum of Akaike weights based on model selection analysis using 667 corrected Akaike's Information Criteria (AIC). Cutoff is set at 0.8 to explore the most essential variables. 668 MAP, mean annual precipitation; MAT, mean annual temperature; HI, humidity index. The black bold 669 values indicate the cutoff values >0.8. 670 24 671 672 Figure captions: 673 674 Figure 1 Geographical distribution of the studied sites included in this meta-analysis. The 675 subset figure indicated the whole range of mean annual precipitation (MAP) and mean annual 676 temperature (MAT) of the study sites and the respective ecosystems. 677 678 Figure 2 Responses of C (a), N (b), P (c) concentrations, C:N (d), C:P (e), N:P (f) ratios in 679 leaves, shoots, litter, soil, and microbial biomass to altered precipitation. Values are weighted 680 effect sizes and their 95% confidence intervals. Values represent the strength of the effect of 681 altered precipitation on response variables relative to the control, numbers indicate the number 682 of observation data. The effect of precipitation treatments is considered significant if the 95% 683 CI of the effect size does not overlap with zero. IPPT, increased precipitation; DP, decreased 684 precipitation. The solid symbols indicate significant response to altered precipitation, and 685 hollow symbols indicate no significant response to altered precipitation. 686 687 Figure 3 Responses of C (a), N (b), P (c) concentrations, C:N (d), C:P (e), N:P (f) ratios in 688 leaves, soil, and microbial biomass to altered precipitation in dry (HI < 0.65) and humid 689 climates (HI > 0.65). Values are weighted effect sizes and their 95% confidence intervals. 690 Values represent the strength of the effect of altered precipitation on response variables relative 691 to the control, numbers indicate the number of data observations. The effect of precipitation 692 treatments is considered significant if the 95% CI of the effect size does not overlap with zero. 693 IPPT, increased precipitation; DPPT, decreased precipitation. The solid symbols indicate 694 significant response to altered precipitation, and hollow symbols indicate no significant 695 response to altered precipitation. 696 697 Figure 4 The sensitivity of C (a), N (b), P (c) concentrations, C:N (d), C:P (e), N:P (f) ratios in 698 leaves, soil and microbial biomass to ΔPPTs binned by slight (≦ 33%), medium (33% ~ 67%) 699 and extreme (≥ 67%) IPPT and DPPT. The error bars represent 95% confidence intervals; the 700 numbers above the error bars represent the observation numbers for each estimate; and the 701 following asterisks indicate significant responses (the 95% confidence intervals do not overlap 702 with zero). IPPT, increased precipitation; DPPT, decreased precipitation. The solid symbols 703 indicate significant response to altered precipitation, and hollow symbols indicate no significant 704 response to altered precipitation. 705 706 Figure 5 A conceptual framework for the double asymmetric responses of plant, soil and 707 microbial biomass C:N:P stoichiometry to altered precipitation in the globe. The solid lines 708 indicate significant response to altered precipitation, and hollow lines indicate no significant 709 response to altered precipitation. 710 711 Figure 6 Potential mechanisms of the influences of altered precipitation on ecosystem 712 processes driving plant, soil and microbial C:N:P stoichiometry in the globe. WUE, water use 713 efficiency; DOC, dissolved organic carbon; NH4+–N, ammonia nitrogen concentration, and 714 NO3—N, nitrate nitrogen concentration. The blue short arrows (↑↓) represent increase and 715 decrease in response to IPPT, red short arrows (↑↓) represent increase and decrease in response 716 to DPPT. ①Wu et al., (2011), ②Zhou et al. (2016), ③Deng et al. (2021). 717 718 Figure 7 Relationships between plant, microbial and soil C:N:P stoichiometry by altered 719 precipitation. S represents the slope of the regression between plant or microbial communities 720 (C:N, C:P or N:P) and resource stoichiometry (soil C:N, C:P, or N:P). Plant or soil microbial 25 721 communities are strong stoichiometric homeostasis whhen the S was not significant (p > 0.05). 722 The data with significant regressions were classified as homeostatic plastic (S > 0.75), weakly 723 plastic (0.5 < S < 0.75), weakly homeostatic (0.25 < S < 0.5), or homeostasis (0 < S< 0.25) 724 (Sterner, & Elser, 2002). 725 726 26 727 728 729 Figure 1 Geographical distribution of the studied sites included in this meta-analysis. The 730 subset figure indicated the whole range of mean annual precipitation (MAP) and mean annual 731 temperature (MAT) of the study sites and the respective ecosystems. 732 733 27 734 735 Figure 2 Responses of C (a), N (b), P (c) concentrations, C:N (d), C:P (e), N:P (f) ratios in 736 leaves, shoots, litter, soil, and microbial biomass to altered precipitation. Values are weighted 737 effect sizes and their 95% confidence intervals. Values represent the strength of the effect of 738 altered precipitation on response variables relative to the control, numbers indicate the number 739 of observation data. The effect of precipitation treatments is considered significant if the 95% 740 CI of the effect size does not overlap with zero. IPPT, increased precipitation; DPPT, decreased 741 precipitation. The solid symbols indicate significant response to altered precipitation, and 742 hollow symbols indicate no significant response to altered precipitation. 743 744 28 745 746 Figure 3 Responses of C (a), N (b), P (c) concentrations, C:N (d), C:P (e), N:P (f) ratios in 747 leaves, soil, and microbial biomass to altered precipitation in dry (HI < 0.65) and humid 748 climates (HI > 0.65). Values are weighted effect sizes and their 95% confidence intervals. 749 Values represent the strength of the effect of altered precipitation on response variables relative 750 to the control, numbers indicate the number of data observations. The effect of precipitation 751 treatments is considered significant if the 95% CI of the effect size does not overlap with 752 zero.IPPT, increased precipitation; DPPT, decreased precipitation. The solid symbols indicate 753 significant response to altered precipitation, and hollow symbols indicate no significant 754 response to altered precipitation. 755 756 29 757 758 759 Figure 4 The sensitivity of C (a), N (b), P (c) concentrations, C:N (d), C:P (e), N:P (f) ratios in 760 leaves, soil and microbial biomass to ΔPPTs binned by slight (≦ 33%), medium (33% ~ 67%) 761 and extreme (≥ 67%) IPPT and DPPT. The error bars represent 95% confidence intervals; the 762 numbers above the error bars represent the observation numbers for each estimate; and the 763 following asterisks indicate significant responses (the 95% confidence intervals do not overlap 764 with zero). IPPT, increased precipitation; DPPT, decreased precipitation. The solid symbols 765 indicate significant response to altered precipitation, and hollow symbols indicate no significant 766 response to altered precipitation. 767 768 30 769 770 771 772 Figure 5 A conceptual framework for the double asymmetric responses of plant, soil and 773 microbial biomass C:N:P stoichiometry to altered precipitation in the globe. The solid lines 774 indicate significant response to altered precipitation, and hollow lines indicate no significant 775 response to altered precipitation. 776 777 31 778 779 Figure 6 Potential mechanisms of the influences of altered precipitation on ecosystem 780 processes driving plant, soil and microbial C:N:P stoichiometry in the globe. WUE, water use 781 efficiency; DOC, dissolved organic carbon; NH4+–N, ammonia nitrogen concentration, and 782 NO3—N, nitrate nitrogen concentration. The blue short arrows (↑↓) represent increase and 783 decrease in response to IPPT, red short arrows (↑↓) represent increase and decrease in response 784 to DPPT. ①Wu et al., (2011), ②Zhou et al. (2016), ③Deng et al. (2021). 785 786 787 32 788 789 Figure 7 Relationships between plant, microbial and soil C:N:P stoichiometry by altered 790 precipitation. S represents the slope of the regression between plant or microbial communities 791 (C:N, C:P or N:P) and resource stoichiometry (soil C:N, C:P, or N:P) responses ratio (RR). 792 Plant or soil microbial communities are strong stoichiometric homeostasis whhen the S was not 793 significant (p > 0.05). The data with significant regressions were classified as homeostatic 794 plastic (S > 0.75), weakly plastic (0.5 < S < 0.75), weakly homeostatic (0.25 < S < 0.5), or 795 homeostasis (0 < S< 0.25) (Sterner, & Elser, 2002). 796 797 33