Grazed grasslands account for a staggering 77% of Earth’s agricultural land (Ellis et al. 2010; Maestre et al. 2022) and more than a quarter of the Earth’s terrestrial ice-free land (Ellis and Mehrabi 2019). This widespread land management practice plays a crucial role in modifying grassland ecosystems and is capable of substantially altering ecosystem carbon (C) dynamics, particularly through interacting with key global change factors (Bardgett et al. 1998; Holland et al. 1996; Miller and Rose 1992). For example, grasslands worldwide are simultaneously being impacted by climate extremes such as drought, which are expected to increase in frequency and intensity with climate change (IPCC 2023). Defoliation is a key element of grazing that has been identified as playing a dominant role in explaining grazer effects on plant and soil properties in grassland ecosystems (Liu et al. 2015; Mikola et al. 2009). As such, it is crucial to understand how defoliation—a proxy for grazing that does not account for factors like trampling and nutrient deposition via animal excreta—affects below-ground C allocation in response to drought (Bardgett and Wardle 2003; Van Der Wal et al. 2004). Such understanding is vital for the development of sustainable land management strategies that reflect the intricacies of grazing dynamics in the real world.

Grassland soil is one of the most significant sinks for atmospheric CO2 globally, but its efficacy depends on management practices and related changes in defoliation intensity caused by ungulate grazing. Generally, long-term, high intensity defoliation can detrimentally impact soil C storage by reducing below-ground C allocation and C supply to soil (Bardgett et al. 2001; Klumpp et al. 2009). In contrast, in the short-term, defoliation can stimulate the translocation of recently assimilated C below-ground, relative to the total uptake, due to alterations in root exudation patterns (Bardgett et al. 1998; Bokhari 1977; Holland et al. 1996; Mawdsley and Bardgett 1997). Previous 13C labelling studies have shown that defoliation significantly increases below-ground 13C allocation (Wilson et al. 2018), and has a positive legacy effect on the translocation of newly fixed 13C to roots (Ma et al. 2021). Defoliation intensity can also alter C flow in grasslands: a meta-analysis of 47 experiments revealed that increased grazing intensity reduced soil C by 18% in C3-dominated grasslands (McSherry and Ritchie 2013). Furthermore, a 13C labelling study found that higher grazing intensity led to decreased 13C in roots and reduced arbuscular mycorrhizal (AM) fungal extraradical hyphal density (Faghihinia et al. 2023). This observation is particularly crucial, given that AM fungi form mutualistic relationships with around 80% of plant species (Smith and Read 2008; Van Der Heijden et al. 1998), and their extraradical hyphal networks not only serve as a vital conduits for C flow from plants to the atmosphere (Johnson et al. 2002a), but also contribute substantially to the global soil C pool (Hawkins et al. 2023). Lastly, prior research suggested that tracking the C flow might offer a more accurate method to study the impact of defoliation on AM fungi (Van der Heyde et al. 2019). Therefore, understanding how defoliation modifies below-ground C allocation in response to drought is critical for developing sustainable land management strategies.

The transfer of recently assimilated C from photosynthesis to soil respiration connects the two most substantial CO2 fluxes in our ecosystem, indicating that even slight modifications in soil respiration can significantly impact C dynamics (Bahn et al. 2009). Perturbations including defoliation can alter soil respiration; recent studies demonstrated that defoliation associated with herbivory can stimulate root and soil respiration (Bai et al. 2021; Holland et al. 1996). But the impacts of defoliation need to be considered alongside key consequences of climate change, notably the predicted increases in duration, intensity, and severity of drought (IPCC 2023; Samaniego et al. 2018). Decades of research have shown that water deficit caused by drought can result in reduced C uptake, and limit subsequent allocation to plant roots and soil (Palta and Gregory 1997). In grassland systems, a history of drought has been found to significantly reduce the uptake of recently assimilated C in plant shoot (Fuchslueger et al. 2016). Additionally, C that was recently fixed through photosynthesis is a critical pathway of energy into soil food webs that drives a multitude of ecosystem processes (Chomel et al. 2022; Leake et al. 2006). This pathway is highly dynamic, and recent photosynthate can be recovered in below-ground pools and fluxes within a few hours of fixation (Johnson et al. 2002a, 2002b; Leake et al. 2006). However, drought impacts on recent photosynthate fluxes are mixed; some studies indicate increased below-ground allocation of newly assimilated C relative to total amount uptake (Barthel et al. 2011; Burri et al. 2014), while others show that drought reduces 13C uptake and root respiration without affecting below-ground allocation (Hasibeder et al. 2015). In contrast, a study of grassland plants suggested that drought shifts more 13C to soil and less to roots, possibly favouring extraradical AM fungal growth (Wang et al. 2021). Given that recent photosynthate is a major pathway of energy that drives myriad soil processes, understanding how defoliation and drought together impact this specific pathway is particularly pressing.

Defoliation and drought can impact AM fungal C storage mechanisms, as reflected by greater concentrations and allocation of recent plant photosynthate to neutral lipid fatty acids (NLFA) (Karlowsky et al. 2018; Mackie et al. 2019). Moreover, drought intensity has been shown to reduce the transfer of recently assimilated C to the energy storage lipid of AM fungi (Oram et al. 2023), and it is well established that drought can reduce soil and root respiration of recently assimilated C (Ingrisch et al. 2020; Meeran et al. 2021; Reinthaler et al. 2021). Yet, how defoliation modifies below-ground C allocation in response to drought, particularly to roots and AM fungal extraradical hyphal networks, remains largely unexplored despite its importance for soil biota and ecosystem functioning. Indeed, while the separate effects of defoliation and drought on C dynamics in grasslands are well documented (Hafner et al. 2012; Hasibeder et al. 2015; Ingrisch et al. 2020; Wilson et al. 2018), their combined effects on the distribution of recent plant photosynthates are poorly understood. As such, understanding how these perturbations simultaneously affect below-ground allocation of newly assimilated C represents an important gap in understanding with respect to how grassland management and climate extremes combine to ecosystem C dynamics.

To address this gap, we executed a fully factorial field experiment in a plant species-rich mesotrophic grassland in northern England, where we manipulated defoliation intensity and subjected these treatments to a simulated summer drought, using rain-out shelters. After the drought, we undertook a 13C pulse-labelling experiment to investigate how below-ground allocation of recent plant photosynthate responds to defoliation intensity and drought, both individually and in combination. The experimental treatments were previously used to test how defoliation impacts the response of AM fungi to drought, showing that defoliation suppresses AM fungi energy storage in response to drought (Xu et al. 2024). Here, we used the same treatments in the subsequent year combined with 13C pulse-labelling to test the following hypotheses: (1) defoliation increases the allocation of 13C to plant roots, soil, and respiration efflux relative to the amount captured by photosynthesis; (2) drought reduces the absolute 13C transfer to plant roots and soil while not affecting the relative 13C transfer to these pools compared to the initial plant uptake; (3) drought stimulates the transfer of 13C into extraradical hyphal network of AM fungi relative to initial plant uptake, whereas defoliation decreases both the absolute 13C enrichment and its relative transfer compared to the initial plant uptake; and (4) defoliation reduces the below-ground transfer of recently photosynthate 13C under drought (Fig. 1).

Fig. 1
figure 1

Theoretical depiction of defoliation and drought impact on 13C excess (13Cexc) and 13C excess relative to initial uptake (relative 13Cexc) across various C pools. From initial plant shoot uptake, to roots, AM fungi extraradical hyphae, soil, and finally 13C-CO2 respiration efflux. Orange minus signs signify negative impacts, while blue plus signs indicate positive effects

Materials and methods

Experimental site and design

The field experiment took place at Colt Park meadows within the Ingleborough National Nature Reserve in the Yorkshire Dales, northern England (54°11′36.6"N, 2°20′52.8"W) (Smith et al. 2008). The plant community was classified as mesotrophic (MG6, MG7, and subcategories) grassland according to the UK National Vegetation Classification (Hall et al. 2004; Rodwell 1992) and was dominated by the grasses Holcus lanatus and Agrostis capillaris, and the forb Plantago lanceolata, which were relatively uniform in their distribution across all plots. The soil is loamy brown earth (pH 5.9, 7.7% C, 0.75% N) and overlays the limestone bedrock of the Malham series (Cole et al. 2019; De Long et al. 2019; Leff et al. 2018). Before the experimental setup, the site experienced managed grazing by sheep in spring, cattle in autumn, and was harvested for hay annually on July 21st (Sweeney et al. 2021; Xu et al. 2024).

The field experimental site, measuring 36 × 31 m, was enclosed to prevent livestock grazing, and the experiment consisted of a three-by-two fully factorial randomized block design, with five replicate blocks in total of 30 plots. The experiment was set up in 2021 when three pairs of plots in each block were randomly subjected to one of three defoliation treatments, including high intensity (applied every two weeks), low intensity (applied every four weeks), and a non-defoliated control from May 24th to September 27th, as reported by Xu et al (2024). In the subsequent year (i.e. 2022), the same treatments were continued from May 9th to July 17th. In both years, these treatments involved removal of all plant shoot biomass to a height of 4 cm above the soil surface (Medina-Roldán and Bardgett 2011). Subsequently, within each defoliation treatments, a drought treatment was randomly assigned for ten weeks: either ambient conditions (control) or a drought simulation using rain-out shelters, simulating a 100-year drought event at this site (Cole et al. 2019). These rain-out shelters were designed to exclude 100% precipitation, and consisted of a wooden frame topped with transparent, corrugated 0.8 mm PVC sheets (Corolux, UK). Each shelter, measuring 155 cm × 130 cm with a height of 60 cm, was tilted to a 15 degrees angle to facilitate optimal water drainage. Each plot consisted of a 40 × 40 cm sampling zone, surrounded by a 100 × 100 cm guard zone to offset potential edge effects. Within each sampling zone, a 25 cm diameter soil ring was randomly buried into the soil for the forthcoming 13C pules-labelling event. The air temperature was tracked using the Thermochron iButton Temperature Logger (Maxim Integrated, CA, USA), while soil temperature was monitored using a WET-2 sensor (Delta-T, UK).

Soil analysis

Soil samples were obtained immediately after the drought shelters were removed on July 17th 2022. Three soil cores (15 cm depth; 2.4 cm in diameter) were randomly taken from the sampling zone using a Soil Probe Kit (Oakfield, WI, USA). These cores were combined, homogenized, sieved (2 mm), and stored at – 20 °C for further analysis. Subsamples designated for phospholipid-derived fatty acids (PLFA) analysis were stored at – 80 °C.

For the assessment of extractable ammonium (NH4+), 5 g of fresh soil was extracted with 25 ml of 1 M KCl, filtered through Whatman no. 42 filter papers (Whatman, UK) then analysed on a Seal AA3 Segmented Flow autoanalyzer (Mequon, WI, USA). The chloroform fumigation method was used to determine microbial biomass C and N (Brookes et al. 1985). Both fumigated and un-fumigated soil (5 g each) were extracted with 25 ml of 0.5 M K2SO4, filtered through Whatman no. 42 filter papers (Whatman, UK) then analysed on a Seal AA3 Segmented Flow autoanalyzer (Mequon, WI, USA). Microbial biomass C and N were calculated by subtracting fumigated with un-fumigated samples, and kEN values of 0.54 and kEC of 0.35 were used respectively for the calculation (Brookes et al. 1985; Cordero et al. 2023).

To examine the root system, after sieving three 15 cm soil cores were carefully washed, and scanned with WinRHIZO (Regent Instrument, Canada). This provided measurements of the total root length (m g−1) and root diameter (mm). The scanned roots were then oven dried to obtain root biomass. The specific root length (SRL) was derived by dividing the dry root biomass by the total root length (m g−1).

Broad-scale shifts in soil microbial community composition was assessed using PLFA analysis, adapted from the Bligh‐Dyer method. PLFAs were extracted from 0.5 g of freeze-dried soil, with C19:0 1,2-dinonadecanoyl-sn-glycero-3-phosphocholine (Avanti Polar Lipids, VA, USA) add as an internal standard to the Bligh‐Dyer extractant at the beginning of the extraction (Buyer and Sasser 2012). Extracted samples were then processed on an Agilent 7890A gas chromatograph (Agilent Technologies, CA, USA). PLFAs were selected as an indicator of fungal and bacterial biomass and were quantified in nmol g−1 dry soil. Specific biomarkers were utilized to represent fungi (18:2ω6,9), gram-negative bacteria (16:1ω9, cy17:0, 18:1ω7 and cy19:0), gram-positive bacteria (a15:0, i15:0, i16:0, a17:0 and i17:0), general bacteria (15:0 and 17:0), and non-specific biomarkers (18:1ω9, 16:0 and 18:0) (Broadbent et al. 2021; Frostegård et al. 2011; Willers et al. 2015).

Mycorrhizal colonization and extraradical hyphal length

Root colonization by AM fungi was determined microscopically using ink staining methods (Vierheilig et al. 1998; Walker 2005). Observations were conducted with a compound microscope (Leica, Germany) at × 200 magnification. Quantifications of arbuscules, vesicles, and total AM fungal root colonization were done using the magnified intersect method (McGonigle et al. 1990).

Prior to the setup of the drought shelters, two mesh bags were buried in each plot to measure both the extraradical hyphal length and 13C enrichment of AM fungi. The first set comprised 30 mesh bags constructed from nylon mesh with a pore size of 35 μm (Plastok Associates Ltd, UK), each bag contained 40 g of autoclaved sand that was sieved to 2 mm, and were used to measure extraradical hyphal length and 13C enrichment (Johnson et al. 2001; Wallander et al. 2001). The second set of 30 mesh bags served as non-AM hyphae 13C enrichment controls. These bags were made of nylon mesh with a much finer pore size of 0.5 μm (Plastok Associates Ltd, UK), filled with 40 g of autoclaved sand, designed to prevent the growth and passage of any hyphae. By subtracting 13C enrichment in both mesh bag final 13C enrichment were obtained. Hyphae were extracted from the sand using a 25 mm 1.2 μm ME28 membrane filter (Cytiva, UK). After staining with an ink-acetic acid solution, the total extraradical hyphal length was quantified using the grid line intersect method (Alef 1995; Brundrett et al. 1994).

13C pulse labelling

We utilized in situ 13C pulse labelling to investigate the initial 13C uptake by the plant shoots, and its subsequent transfer to roots, extraradical AM fungal hyphae, soil, and back into the atmosphere as 13C-CO2 over a 28-day period (Fig. 1).

One day prior to the labelling event, on July 17th, drought shelters were removed. All plots were then watered with four litres each to alleviate the drought and simulate photosynthetic activity. The pulse labelling took place on July 18th, using 30 airtight chambers constructed from 10 L transparent plastic bell cloches. During the labelling process, each chamber was securely joined to a ring embedded within the respective plot to ensure an airtight seal.

For enhancing air circulation within the canopies, a battery-powered brushless fan was mounted inside each chamber. Additionally, a 50 ml centrifuge tube was affixed inside each chamber to hold 10 ml 3 M H2SO4. The 13CO2 was generated by injecting an aqueous solution of NaH13CO3 (99 atom %) (Sigma-Aldrich, MI, USA) into the H2SO4 (Pang et al. 2021; Stewart and Metherell 1999). A total of 20 ml of the solution, containing 1 g of NaH13CO3, was injected in four separate intervals, each approximately 30 min apart and each injection carrying 5 ml of the solution. The 13CO2 concentration was actively monitored using a Picarro G2201-i Isotopic Analyzer (Picarro Inc., CA, USA) connected to a mobile greenhouse gas laboratory. 13CO2 concentration were maintained below 800 ppm, once the 13CO2 concentration began to decrease towards equilibrium, the subsequent injection was initiated.

Post 13C labelling sampling

Prior to the labelling, a set of plant shoot, root, and soil sample were taken as unlabelled natural abundance controls. Post labelling sampling time points were set to 0 day (immediately after labelling), 2 days (44 h), 4 days, 14 days, and 28 days after the labelling. The 25 cm diameter labelling area was divided into five sections, each corresponded a different sampling point.

Shoots were clipped within each section of the soil ring and root and soil samples were obtained from two soil cores, which were then pooled, homogenized, and passed through a 2 mm sieve. The remaining root samples were washed with water. All holes created by the soil core were filled with 50 ml centrifuge tubes to minimise potential gas exchange from the exposed surface and to mitigate disturbance. 13C-CO2 respiration efflux was measured at every sampling interval using a dark airtight chamber connected to Picarro G2201-i Isotopic Analyzer (Picarro Inc., CA, USA). The 13C:12C isotopic ratio inside the chamber were continuously measured.

Previous in situ pulse labelling experiments show that peak 13C enrichment in hyphal respiration occurred 34 h (Johnson et al. 2002a) and in soil colonized by AM fungi occurred 70 h (Johnson et al. 2002b) after the labelling period. Therefore, to avoid missing maximum outputs of 13C from roots and AM fungi, two sets of mesh bags were harvested 2 day (44 h) after the labelling, and by calculating the difference in 13C atom percent excess between the 35 μm and 0.5 μm mesh bags, we derived the 13C enrichment of AM fungi extraradical hyphal in the mesh bags. All samples were immediately placed in an oven and dried at 70 °C. The shoot and root biomass, soil bulk density was determined after the final harvest and were ground, then analysed using a Picarro Combustion Module Cavity Ring-Down Spectroscopy (Picarro Inc., CA, USA).

13C calculations

All 13C data were converted from δ13C to 13C atom % against international standards and the 13C:12C isotopic ratio. By subtracting the 13C atom percentage from the unlabelled controls, 13C atom % excess for all data sets were obtained (Johnson et al. 2011). To calculate the total amount of 13C (mg) incorporated into different pools from the 13C pulse-labelling, the 13C atom % excess was integrated with the total C content in each pool (Chomel et al. 2022; Johnson et al. 2011; Meeran et al. 2021). Although all plots were labelled with the same amount of 13C-CO2 simultaneously, variations in plant shoot biomass and photosynthetic rates resulted in differing amount of 13C initially taken up by plants. Hence, we used 13C transfer in different pools relative to the initial 13C uptake by plant shoots to compare 13C transfers among treatments (Chomel et al. 2022).

Data analysis

Statistical analyses were carried out using R version 4.2.0 (R Core Team 2023). All data were examined for normality and log-transformed or square root transformed (only for data containing zero value) if necessary.

The effects of drought and defoliation were analysed using mixed-effects model. The function ‘lmer’ from the LME4 package was used to construct the model. Defoliation and drought treatments were set as fixed effects, with a random effect of blocks. We assumed an interaction between defoliation and drought treatment for all response variables, but if not, the interaction was removed from the model.

The function ‘lme’ from the NLME package was used to analyse the impacts of defoliation and drought on the 13C enrichment and its relative transfer among different pools. We used ‘lme’ function because it allows for different variances for each post sampling time point. The model was formulated as follows:

$$lme(response \sim defoliation \times drought, random= \sim 1|Block, weights=varIdent \left(form= \sim 1|time point\right), data=data)$$

All results from the mixed-effects model were obtained by ANOVA and followed by a Tukey Honestly Significance Difference Test (Tukey HSD) among different defoliation treatment levels and interactions.

To test the drought impacts, and its interaction with different defoliation intensities, we calculated a log Response Ratio (logRR) of the 13C enrichment in each pool from each pair of drought and ambient control plot. An effect size of 0 indicates no effect, a positive value signifies a positive effect, while a negative logRR denotes a negative impact of the drought (Chomel et al. 2022). The same mixed-effect models, utilizing the ‘lme’ function were used to analyse the defoliation responses to drought. Last, the function ‘emmeans’ from EMMEANS package was used to calculated 95% confidence interval of predicted means and were used to specify if the logRR differed significantly from 0.


The drought treatment significantly reduced soil moisture by an average of 23.5% during the ten-week drought period (P < 0.001, F = 591.2; Fig. S1). Additionally, drought treatments had a significant negative legacy effect on soil moisture, with soil moisture being on average 5% lower in droughted than non-droughted soils during the post-drought, labelling period (P < 0.01, F = 7.383; Fig. S2).

Effects of defoliation and drought on plant properties and AM fungi symbiosis

Drought significantly reduced total root colonization by AM fungi by an average of 29.6% (P < 0.001, F = 31.34; Fig. 2a), colonization of arbuscules by an average of 51.9% (P < 0.001, F = 32.89; Fig. 2b), and extraradical hyphal length by an average of 26.3% (P < 0.01, F = 7.959; Fig. 2c). Defoliation also significantly reduced colonization of arbuscules (P < 0.05, F = 3.559; Fig. 2b), with high intensity defoliation reduced colonization of arbuscules compared to the control by average of 15.7%. Moreover, defoliation significantly decreased extraradical hyphal length (P < 0.001, F = 13.46; Fig. 2c), results from the Tukey HSD test indicated that high intensity defoliation significantly decreased the length of hyphae compared to the non-defoliated control (P < 0.001), by an average of 52.4%.

Fig. 2
figure 2

Responses of AM fungal (a) total root colonization (%), (b) colonization of arbuscules (%), (c) extraradical hyphal length (m g−1), and (d) total root length (mg g−1), (e) specific root length (mg g−1), and (f) root diameter (mm) to defoliation and drought. Boxplots present the mean, outliers, SE and range, within each plot, values with the same letter are not significantly different. Asterisks highlight the significant effects of defoliation, drought, and their interactions (* P < 0.05, ** P < 0.01, *** P < 0.001)

Defoliation had a positive effect on total root length immediately after the drought (P < 0.05, F = 3.480; Fig. 2d). There was a significant interaction between drought and defoliation intensity (P < 0.05), with low intensity defoliation increasing total root length under drought conditions. By contrast, the combination of defoliation and drought negatively affected root diameter immediately following drought (P < 0.01, F = 6.411; Fig. 2f), with low intensity defoliation and drought specifically suppressed root diameter (P < 0.05). Furthermore, drought caused a significant decline in SRL immediately after drought (P < 0.05, F = 4.316; Fig. 2e).

Defoliation significantly decreased soil concentrations of extractable NH4+ (P < 0.01, F = 6.229; Fig. 3b), with both high and low intensity defoliation treatments significantly reducing this measure compared to the non-defoliated control (P < 0.05, P < 0.05; Fig. 3b). Drought exerted a significant negative impact on soil total PLFA (P < 0.05, F = 5.112; Fig. 3a), and microbial C and N immediately after drought (P < 0.001, F = 29.82; Fig. 3c; P < 0.01, F = 13.11; Fig. 3d). Lastly, drought did not influence shoot biomass at final harvest (Table. S4).

Fig. 3
figure 3

Changes in soil (a) total PLFA (nmol g−1), (b) extractable NH4+ (mg kg−1), (c) microbial C (mg kg−1), and (d) microbial N (mg kg−1) in response to defoliation and drought. Boxplots show the mean, outliers, SE and range. Asterisks denote the significant impacts of defoliation and drought treatments (* P < 0.05, ** P < 0.01, *** P < 0.001)

Effects of defoliation on C allocation in response to drought

Drought reduced 13C enrichment in plant shoots, roots, soil, and 13C-CO2 respiration efflux compared to the ambient control. However, the effects of drought were not seen under the defoliation treatments, where drought had no detectable effect on 13C uptake of shoot, its transfer to roots, extraradical AM fungal hyphae, and 13C-CO2 respiration efflux. However, both low and high-intensity defoliation treatments combined with drought led to a significant reduction in relative 13C enrichment of the soil. This observation was supported by findings from the logRR (Fig. 4) and corroborated by the results from the mixed-effect model (Fig. 5), where drought had a more pronounced effect, resulting in reduced relative 13C enrichment in the soil. Drought not only significantly reduced soil 13C enrichment (P < 0.001, F = 21.14; Fig. 5: Table. S2), but its interaction with defoliation further exacerbated such decreased in soil 13C enrichment (P < 0.05, F = 3.104; Fig. 5: Table. S2). The Tukey HSD test results revealed that both high and low intensity defoliation treatments had a significant negative interactive effect with drought on soil 13C enrichment (P < 0.01, P < 0.01; Fig. 5), all suggesting that less 13C was transferred to the soil pool.

Fig. 4
figure 4

Impact of defoliation treatments on post-drought 13C enrichment across various C pools. The response ratio, using the logarithm of the drought to control ratio (log (drought/control)). The positive or negative sign of the logRR indicates the direction of the drought effect on the 13C enrichment, while a value of zero denotes no significant post-drought effect. The logRR plot displays mean, SE and range. Black asterisks highlight significance from the mixed-effects model for defoliation effects on logRR (* P < 0.05, ** P < 0.01, *** P < 0.001). Red asterisks indicate significance of drought effect according to the confidence intervals of predicted means from the mixed-effects model (Table. S1)

Fig. 5
figure 5

Impact of defoliation and drought treatments on 13C enrichment of bulk soil. The dashed and soil lines represent drought and ambient (control) treatment. Different colours signify non-defoliated control, low, and high intensity defoliation treatments. Asterisks indicate significant effects of drought and interaction between defoliation and drought treatments. Treatment combination sharing the same letter are not significantly different. (* P < 0.05, ** P < 0.01, *** P < 0.001)

Effect of defoliation and drought on absolute 13C enrichment

Drought significantly reduced 13C enrichment of plant shoots (P < 0.001, F = 63.69; Fig. 6a). Immediately following the labelling, the average 13C enrichment in shoots was 0.59 atom % under drought conditions and 0.88 atom % in the ambient treatment. A similar trend was found in plant roots, where drought markedly reduced 13C enrichment (P < 0.001, F = 43.48; Fig. 6b). In response to drought, the peak root 13C enrichment occurred 4 days after the labelling, averaged at 0.038 atom % excess. In contrast, with the ambient treatment, the peak was exhibited 14 days post-labelling, and averaged 0.072 atom % excess.

Fig. 6
figure 6

Effects of drought treatments on the 13C enrichment of (a) plant shoots, (b) roots, (c) soil, and (d) 13C-CO2 respiration efflux, measured in atom % excess. Different colours correspond to control ambient and drought treatments. Significant drought treatment effects are denoted by asterisks. (* P < 0.05, ** P < 0.01, *** P < 0.001)

The drought treatment also reduced soil 13C enrichment as previously mentioned above (P < 0.001, F = 21.14; Fig. 6c). The apex of soil 13C enrichment was observed 28 days after the labelling, averaging 0.0017 for drought and 0.0032 atom % excess with ambient treatments. Furthermore, drought significantly reduced 13C-CO2 respiration efflux (P < 0.001, F = 35.00; Fig. 6d). The maximum 13C-CO2 respiration efflux was found a day after the labelling, averaging 0.0024 with drought and 0.0062 atom % excess with the ambient treatment. Overall, defoliation treatments had no detectable effect on plant shoots, roots, soil, and respiration efflux 13C enrichment.

While drought did not influence 13C enrichment in AM fungi extraradical hyphae, defoliation significantly reduced extraradical AM fungal hyphae 13C enrichment a day after the labelling (P < 0.001, F = 12.54; Fig. 7a). The high intensity defoliation treatment significantly reduced 13C enrichment in the AM fungi extraradical hyphae compared to the non-defoliated control (P < 0.001, F = 12.54; Fig. 7a).

Fig. 7
figure 7

Impacts of defoliation and drought treatments on AM fungal extraradical hyphae (a) 13C enrichment one day after the labelling, and (b) 13C transfer relative to the initial plant shoot uptake. Distinct colour indicates control ambient and drought treatments. Asterisks denote significant from defoliation and drought treatments. (* P < 0.05, ** P < 0.01, *** P < 0.001)

Effect of defoliation and drought on 13C allocation relative to plant uptake

Both defoliation and drought treatments individually and notably reduced the initial 13C uptake in plant shoots (P < 0.001, F = 51.69; P < 0.001, F = 75.17; Fig. 8a). Both high and low intensity defoliation significantly (P < 0.001) impacted the initial 13C uptake (Fig. 8a). Defoliation significantly increased the 13C transfer relative to plant uptake to roots (P < 0.001, F = 12.51; Fig. 8b) with both high and low defoliation intensity significantly stimulating relative 13C transfer to roots compared to the control (P < 0.001). Likewise, defoliation treatments increased 13C transfer relative to plant uptake to soil (P < 0.001, F = 26.71; Fig. 8c) with both high and low intensity defoliation treatments having a significant positive effect on the relative 13C transfer to soil compared to the non-defoliated control (P < 0.001). Similar enhancement was also found in the 13C transfer relative to plant uptake via the 13C-CO2 respiration efflux (P < 0.01, F = 5.152; Fig. 8d), whereby high intensity defoliation treatments significantly enhanced the relative 13C transfer via respiration efflux (P < 0.01). In contrast, the drought treatment significant reduced relative 13C transfer via respiration efflux (P < 0.05, F = 6.792; Fig. 8d).

Fig. 8
figure 8

Effects of defoliation and drought treatments on the 13C uptake of (a) plant shoots, and 13C transfer relative to uptake, of (b) roots, (c) soil, and (d) 13C-CO2 respiration efflux. Colours distinguish between non-defoliated control, low, and high intensity defoliation treatments. Solid lines represent ambient treatment, while dashed lines indicate drought treatments. Asterisks mark significant effect from defoliation and drought treatments. (* P < 0.05, ** P < 0.01, *** P < 0.001)

Overall, the defoliation treatments significantly suppressed the 13C transfer relative to plant uptake to extraradical AM fungal hyphae (P < 0.05, F = 4.582; Fig. 7b), with high intensity defoliation having the greatest effect (P < 0.05). The drought treatment significantly increased the relative 13C transfer to extraradical AM fungal hyphae (P < 0.05, F = 7.004; Fig. 7b).


Our findings support the first hypothesis that defoliation not only simulates the allocation of recent assimilated 13C transfer to plant roots and soil, but also that high intensity defoliation enhances 13C-CO2 efflux relative to plant uptake, suggesting a potential loss of recently photosynthesized C via soil respiration. Our results also corroborate our second hypothesis that drought reduces 13C uptake by plant shoots and diminishes the 13C enrichment in both roots and soil. Our findings are also consistent with our third hypothesis that drought alone stimulates the transfer of 13C into fungal extraradical hyphae relative to initial plant uptake, while defoliation diminishes both the absolute transfer of 13C and its transfer relative to initial uptake. Lastly, our study supports our fourth hypothesis that defoliation reduces below-ground C allocation in response to drought, evidenced by lower soil 13C enrichment compared to the non-defoliated control, potentially impacting soil C dynamics and other ecosystem processes (Klumpp et al. 2009).

Despite the defoliation treatments reducing the amount of 13C uptake by plant shoots (in accordance with plant biomass removal), it did stimulate below-ground allocation of 13C to roots and soil relative to initial plant uptake. Our findings are consistent with earlier studies showing that both insect herbivory and simulated ungulate defoliation can increase the short-term flow of recent assimilate below-ground (Doll 1991; Holland et al. 1996). Such a response may be attributed to alterations in root exudation patterns (Bardgett et al. 1998), as shown experimentally for grassland plants using 13C pulse-labelling (Hamilton et al. 2008). Additionally, simulated defoliation has been shown to induce the production of lignin-rich roots that rely on recent assimilate, which also aligns with our findings (Ziter and MacDougall 2013).

As expected, we found that drought alone significantly reduced 13C uptake of plant shoots, along with the 13C enrichment in root, soil, and respiration efflux. These findings align with previous drought studies across grasslands in multiple locations, which have shown that drought decreases total C uptake by plants and can lead to a reduction of above-ground C storage (Burri et al. 2014; Chomel et al. 2022; Fuchslueger et al. 2014; Ingrisch et al. 2020; Karlowsky et al. 2018). However, we found no impact of drought on the allocation of 13C to below-ground relative to the amount initially fixed by plants, which corroborates our second hypothesis. Consideration of past work (Burri et al. 2014; Chomel et al. 2019; Hasibeder et al. 2015; Karlowsky et al. 2018; Sanaullah et al. 2012) indicates that drought can have contrasting effects on below-ground C allocation (relative to the amount fixed by plants). Such disparities may arise from differing drought intensities applied to experimental plots, with previous work showing that when drought intensity exceeds 75% soil water deficit, 13C allocation to NLFA declines sharply (Oram et al. 2023). Another factor that may contribute to variations in below-ground C allocation under drought conditions are differences in the strategies of dominant plant species (Ingrisch et al. 2020). Although unlikely in the short timeframe of this experiment, longer-term defoliation and drought treatments may lead to shifts in the relative abundance of plant species and their roots (Smith et al. 2003). Nevertheless, our findings demonstrate the potential for drought in combination with defoliation to impact the allocation of recently assimilated C below-ground in the short-term, which could potentially have indirect implications on soil C dynamics.

Few studies have attempted to quantify 13C transfer to extra-radical AM fungi in the field, and so our analysis provides valuable data on this key process. Carbon transfer to mycorrhizal fungi is critical in driving their activity and thus the capacity to acquire growth limiting nutrients from soil (Johnson et al. 2001). While drought did not impact 13C enrichment in extraradical AM fungal hyphae, it significantly stimulated 13C transfer relative to the plant uptake. This response aligns with prior studies, which suggested that plants enhance C allocation into AM fungal energy storage lipids under drought conditions (Karlowsky et al. 2018; Mackie et al. 2019). Supporting this view, AM fungi are capable of enhancing resistance of plants to drought (Augé, 2001; Jia et al. 2021; Li et al. 2013; Mariotte et al. 2017; Remke et al. 2021), which is consistent with our finding that drought did not affect above-ground biomass production.

Moreover, compared with low intensity treatment, high intensity defoliation significantly diminished both the 13C enrichment of AM fungi extraradical hyphae and its 13C allocation relative to plant uptake. Drought also led to a significant decline in AM fungi root colonization, and both drought and defoliation individually resulted in reduced colonization of arbuscules, and a decrease in extraradical hyphal length. Despite drought significantly reducing 13C enrichment in all pools, it had no apparent effect on AM fungal extraradical hyphal length. This might be attributed to the constraints imposed by the size of the labelling chambers which meant that we were limited to one set of samples for analysis of hyphal 13C enrichment. When these data were compared with soil 13C enrichment taken 2 days post-labelling (44 h), we observed a similar non-significant drought effect (Fig. S3).

Collectively, our findings support the C limitation hypothesis (Gehring and Whitham 2002). AM fungi are heavily dependent on photosynthates from their host plant (Johnson et al. 2006), and removal of leaf tissue can reduce photosynthate production, subsequently hampering the ability of plants to support AM fungal partners (Bethlenfalvay and Dakessian 1984; Gehring and Whitham 2002). Interestingly, we did not observe a decrease in total root colonization by AM fungi, but this observation supports a previous meta-analysis, across 99 studies, which found herbivory typically reduced colonization by only 3% (Barto and Rillig 2010). These results are also supported by real-world grazing experiments, which show that grazing generally has no effect on root colonization unless it involves heavy grazing intensities (Ba et al. 2012; Cavagnaro et al. 2019; Van der Heyde et al. 2017).

The mutualistic relationship between plants and AM fungi is often influenced by C allocation. Recent research has shown that under insect defoliation, plants prioritize C allocation to AM fungi over parasitic nematodes, achieving this through the maintenance of fatty acid biosynthesis and transportation pathways (Bell et al. 2024). In contrast, our findings underscore that although drought supresses total 13C uptake by plants, it increases the allocation of recently photosynthesized C to AM fungi. This investment likely represents an adaptive strategy to mitigate the impact of drought on above-ground biomass production. Such dynamics emphasizes the essential mutualistic relationship between AM fungi and their host plants under drought conditions. However, high intensity defoliation seems to undermine such beneficial plant-fungi interaction, supporting our third hypothesis.

Our analysis also revealed that drought had multifaceted effects on root and AM fungal parameters. It not only led to a decline in SRL and increase in root diameter, but also impacted AM fungal root colonization and extraradical hyphal length. Our observations partially resonate with earlier research where drought was seen to reduce SRL without affecting the extraradical hyphal length (Miller et al. 1995). Variability in how AM fungal root colonization and extraradical hyphae respond to drought might be attributed to the differing intensities of drought or the occurrence of recurrent drought events. Past research indicates that recurrent droughts can significantly alter the AM fungal community (Canarini et al. 2021) and colonization (Jentsch et al. 2011). These findings support the idea that repeated or prolonged stress from drought can have lasting impacts on plant-fungal symbiotic relationships. Notably, even though drought suppressed AM fungi root colonization and the growth of extraradical hyphae, our findings show that plants still invested large a proportion of recently assimilated C below-ground, with a particular emphasis on AM fungi extraradical hyphae. This highlights the crucial role of plant-fungi interactions under drought conditions.

Our study provides new insights into the combined effects of defoliation and drought on the allocation of recently assimilated C in soil in the field, and supports our last hypothesis that defoliation reduce 13C enrichment below-ground in response to drought. The balance between C assimilation and C loss through respiration efflux can reflect rates of C accumulation in soil (De Deyn et al. 2008). Hence, our observation suggests that, in the short term, defoliation and drought can potentially adversely affect below-ground C flux in a temperate grassland. Furthermore, we discovered that while drought supresses 13C transfer relative to plant uptake through 13C-CO2 respiration efflux, high intensity defoliation significantly increases it compared to the non-defoliated control. This observation aligns with previous research (Bai et al. 2021; Holland et al. 1996), and demonstrates that while drought limits 13C-CO2 respiration relative to plant uptake, high intensity defoliation enhances it.


Our findings highlight that when combined defoliation and drought negatively affect the allocation of recently assimilated C below-ground to soil. We also demonstrate that high intensity defoliation amplifies the transfer of 13C via respiration efflux relative to the uptake, potentially leading to a significant loss of C. Furthermore, drought stimulates the relative 13C transfer into AM fungal extraradical hyphae, whereas high intensity defoliation reduces its relative transfer, weakening the vital plant-fungi mutualistic relationship. Additionally, defoliation treatments significantly stimulate the allocation of 13C belowground relative to initial uptake. Both the frequency and intensity of drought are expected to increase in the future (IPCC 2023), and factors such as grazing, rainfall seasonality, and mean annual precipitation are significant predictors in regulating soil C storage (Maestre et al. 2022). Therefore, our findings emphasize that defoliation can modify how future droughts impact soil C dynamics with potential impacts on soil C storage. Given the pivotal role of the extraradical hyphae network of AM fungal as a key pathway to the soil organic matter pool (Frey 2019; Godbold et al. 2006; Huang et al. 2021; Rillig et al. 2001), and recognizing that soil respiration represents one of the major terrestrial C flux linking the soil with the atmosphere (Bond-Lamberty and Thomson 2010). Our findings advocate for the development of resilient grassland management practices, and specifically the avoidance of high intensity grazing or mowing to minimize the loss of recent assimilate C through CO2 respiration efflux, and safeguard the critical plant-AM fungal mutualistic relationship.