Snowfall Change Had Different Effects on Litter Decomposition for Two Typical Desert Species in Different Periods

Abstract

In desert ecosystems, litter decomposition is the primary source of soil nutrients and is strongly affected by extreme climate events, which may influence desert plant survival and species diversity. To date, the effects of snowfall changes on litter decomposition in desert species remain poorly understood. Here, a snowfall manipulation experiment was conducted in Northwest China that included snowfall addition and removal treatments, as well as a natural snowfall control. Compared to the control, snowfall addition increased the amount of litter mass lost for Salsola passerina and Reaumuria soongarica during the snow-covered period by 21.54% and 21.8%, respectively. In contrast, snowfall addition effects differed between species during the snow-free period. More carbon was released from the S. passerina litter in the snowfall addition treatment during the snow-free period. Similarly, during the snow-covered period, more carbon and nitrogen were released from the R. soongorica litter in the snowfall addition treatment. Overall, the proportion of litter mass lost (from the annual total) increased with snowfall addition in the snow-covered period but was reduced with snowfall addition in the snow-free period. In the snow-covered period, the snowfall addition treatment affected litter mass loss to the same extent in both species but impacted S. passerina more strongly than R. soongorica in the snow-free period due to differences in soil urease activity. Changes in snowfall, therefore, significantly influenced litter decomposition in both desert species, but these effects differed between the snow-covered and snow-free period, particularly for litter with a higher C:N ratio.

1. Introduction

Over the past half a century, global warming has caused more precipitation to occur in the form of rainfall [1,2]. Temperature increases have also accelerated the global water cycle, which has resulted in both higher precipitation and extreme snowfall in the winter [3,4]. Compared to rainfall, snowfall exhibits a more complex response to climate change, with greater spatiotemporal variability, making climate change impacts on snowfall a current research hotspot. Snowfall is a key component of the hydrological system in the arid region of Northwest China, a region which is also extremely sensitive to climate change. Previous studies have shown significant increases in the annual snowfall in this region, also revealing that extreme snowfall events have increased in frequency [5], greatly impacting biogeochemical cycling.

Litter decomposition is a critically important ecological process in desert ecosystems. It not only provides essential nutrients for desert plants, as well as both nutrients and energy to soil microorganisms, but also plays an indispensable role in desert carbon and nitrogen cycling [6,7]. Litter decomposition is affected by many factors, such as the local climate, litter quality, and soil conditions [8]. In desert ecosystems, the vegetation is the most sensitive component to rainfall change. In addition to growing season rainfall, winter snowfall also plays a vital role in ecosystem functioning in desert ecosystems. The amount of snowfall can affect the light availability, soil moisture, and temperature, thereby also impacting the activity of soil organisms and the rate of litter decomposition. An earlier study found that litter decomposition was arrested in winter, but other studies have shown that greater snow cover can accelerate litter decomposition and the associated release and transformation of soil carbon and nitrogen [9,10,11]. However, the ecological processes and mechanisms underlying snow cover effects on litter decomposition in desert ecosystems remain unclear [12].

Salsola passerina Bunge and Reaumuria soongarica (Pall.) Maxim. are two typical desert plants that are cold and drought resistant, as well as capable of growing on alkaline sandy soils. These two species often occur together in mixed communities. They play an important role in the succession of desert ecosystem communities while stabilizing local environmental conditions [13,14]. Nutrients released from R. soongarica and S. passerina litter help shape soil organic carbon and nutrient dynamics in desert soils. In a previous study, the rate of litter decomposition and nutrient release from both R. soongarica and S. passerina litter increased with precipitation in a desert region [15]. However, how shifts in snowfall may affect litter decomposition in desert ecosystems remains unknown. Therefore, we conducted a snowfall control experiment on the two litters with snowfall removal, natural snowfall and the addition of snowfall in the desert region to test the following hypotheses: (1) the mass loss and nutrient release of litters under snowfall removal (SR) would be lower than that in the snowfall added (SA) treatment during the snow-covered period, and this result is the opposite in the snow-free (SF) period; (2) the effect of snowfall on litter decomposition would differ between R. soongarica and S. passerina due to differences in litter chemistry.

2. Materials and Methods

2.1. Study Area

This study was conducted in the desert ecosystem located in the Linze Inland River Basin of the central Hexi Corridor in Northwest China (39°41′ N, 100°12′ E). In this ecosystem, the mean annual temperature is 7.6 °C. The average annual precipitation measures 117 mm, and the average annual evaporation is 2390 mm. The most common shrubs are Salsola passerina and Reaumuria soongorica, while common herbs are Suaeda glauca, Halogeton arachnoideus, and Eragrostis minor. More information about the desert ecosystem can be found in Xie [15].

2.2. Experimental Design

At the end of October 2022, eighteen replicate plots (2 m × 2 m) were established in the study area with 4–5 m spacing between plots. The plots were randomly divided into three snowfall treatments: natural snowfall (CK), snow removal (SR), and snow addition (SA). Each treatment had six replicates. The snowfall manipulation experiment was conducted from 20 November 2022 to the end of February 2023. The snow removal treatment employed stainless-steel frames covered with transparent acrylic plates (2 m × 2 m × 0.5 m) to exclude snowfall in the SR plots. After each snowfall, the snow collected on top of the acrylic plates covering each SR plot was added to the corresponding SA plots; the snow was sifted through a large-mesh screen to simulate natural snowfall. For each snowfall event, the snow depth was measured at six random locations per plot using a ruler (eight events in total; Figure 1). To reduce the effects of snowmelt on litter decomposition in adjacent plots, each plot was enclosed with polyvinyl chloride (PVC) panels in early March. During the snow-free period (from March to 20 November 2023), the steel frames were removed from all SR plots. The soil temperature was measured using a soil temperature and humidity meter (TZS-2X) at a depth of 5 cm (Figure 1).

2.3. Litter Decomposition and Sampling

Leaf litter from R. soongarica and S. passerina was collected in the study area in October 2022. The litter was air-dried for two weeks; then, 10 g of each litter type was placed into a 10 × 10 cm nylon mesh bag with a mesh size of 0.5 mm. Meanwhile, six litter samples per species were oven-dried at 80 °C to compare the mass of air- versus oven-dried samples. Samples were then ground prior to determination of chemical traits, such as the cellulose, lignin, organic carbon, total nitrogen, and total phosphorus content.

A total of 72 litterbags (two litter types × three treatments × six replicates × two sampling dates) were positioned on the soil surface of the 18 experimental plots in December 2022. To prevent the litterbags from blowing away, all litterbags were secured with iron wire.

To assess how snowfall treatments affected litter decomposition, the experimental period was divided into two parts: a snow-covered (SC) and snow-free (SF) period. In the study area, the snow-covered (SC) period typically lasts from late November to the end of February, while the snow-free (SF) period is from March to November. Therefore, six litterbags per species per treatment were harvested at the end of February and November 2023, respectively; these were immediately transported to the laboratory. The litter samples were oven-dried at 80 °C to a constant weight; after recording the weight, samples were ground and analyzed to determine element concentrations. The total organic C content was quantified using the K₂Cr₂O₇ oxidation method. The total N content was quantified via the Kjeldahl method, while the total phosphorus (P) content was determined using the molybdenum blue colorimetric method [16]. Cellulose and lignin concentrations were measured using the acid-detergent fiber method [17].

2.4. Soil Sampling and Analysis

Surface soils (0–5 cm) were sampled in each plot at each site where a litterbag had been positioned. The soil samples were passed through a 2 mm mesh sieve in the laboratory. Part of each fresh sample was air-dried prior to quantifying the soil organic carbon and total nitrogen concentrations. The soil organic C (SOC) content was measured via the K₂Cr₂O₇ oxidation method, while the soil total N (STN) content was assessed using the Kjeldahl method. The remainder of the fresh soil samples was used to measure soil enzyme activity. The activity of soil cellobiohydrolase (CBH) and peroxidase (POD) enzymes was determined following Luo [18]. The soil alkaline phosphatase (AKP) activity was quantified as described by Sinsabaugh [19], and the soil urease (Ure) activity was measured following Maisto [20].

2.5. Statistical Analysis

The mass loss (ML, %) for each sample and nutrients released (NL, %) from the litterbags were calculated as follows [21]:

$$\mathrm{Mass} \mathrm{loss}(\%)={\displaystyle \frac{M_0-M_t}{M_0}}\times 100$$

(1)

$$\mathrm{Nutrient} \mathrm{release}(\%)={\displaystyle \frac{M_0\times C_0-M_t\times C_t}{M_0\times C_0}}\times 100$$

(2)

where $M_0$ is the initial dry mass (g) and $M_{\mathrm{t}}$ is the dry mass at time t. $C_0$ and $C_t$ represent element concentrations (mg g⁻¹) for initial samples and at later sampling dates (time t), respectively.

The relative contribution of each sample to the annual litter mass loss (Pi, %) was calculated as follows [22]:

$$P_{\mathrm{i}}(\%)={\displaystyle \frac{M_{t+1}-M_t}{M_0-M_T}}\times 100$$

(3)

where ($M_{t+1}-M_t$) is the difference in leaf litter mass between two sampling dates, and $M_T$ is the dry mass (g) at the final sampling date (T).

Initial leaf litter chemical properties were compared between R. soongorica and S. passerina using Student’s t-tests. Snowfall treatment effects on litter mass loss, nutrient release, and soil enzyme activity were assessed using one-way analyses of variance (ANOVAs) (p < 0.05). The relationships between litter decomposition and diverse soil properties were explored using a correlation matrix. Redundancy analysis (RDA) was performed in CANOCO 5.0; a correspondence analysis was carried out prior to the RDA to ensure linear relationships among variables [23]. All statistical analyses were completed in IBM SPSS 26.0 (SPSS Inc., Chicago, IL, USA), and all figures were generated in Origin 2022 (Origin Lab Inc., Chicago, IL, USA).

3. Results

3.1. Initial Chemical Characteristics of the Leaf Litter

The initial chemistry of the leaf litter differed between R. soongarica and S. passerina, with the exception of the lignin content (Figure 2). The R. soongorica litter had greater C, cellulose, N, and P content, but the S. passerina litter had higher C:N, C:P, and lignin/N ratios.

3.2. Litter Mass Loss for the Two Litter Types

Compared to the natural snowfall control (CK), litter mass losses were greater in the SA treatment for both litter types during the snow-covered (SC) and snow-free (SF) periods, with the exception of the R. soongorica litter in the SF period (Figure 3). In the SR treatment, the amount of litter mass lost also declined over the SC period but did not differ from that of CK in the SF period. The decomposition rate (K) of both litter types was highest in SA (K = 0.069 for S. passerina and 0.083 for R. soongorica) and lowest in SR (K = 0.043 for S. passerina and 0.058 for R. soongorica). Across treatments, R. soongorica showed greater litter mass losses than S. passerina during both the SC and SF period, but this difference was only significant during the SF period.

For S. passerina litter, the proportion of the total mass loss that occurred during the SC period measured 48.88% in the SA treatment and 43.31% in the SR treatment. Compared with CK, this represents an 2.5% increase in SA and a 9.2% reduction in SR. In the SF period, the relative mass loss was 8.4% higher in SR and 2.3% lower in SA, as compared to CK (Figure 4A). For R. soongorica litter, the proportion of the total mass loss was 8.8% higher in SA and 9.3% lower in SR than in CK during the SC period; however, it was 6.5% lower in SA and 6.9% higher in SR than in CK during the SF period (Figure 4B).

3.3. Carbon and Nitrogen Released from the Two Leaf Litter Types

Compared with CK, the carbon and nitrogen released from S. passerina litter did not differ in the SR and SA treatments during the SC period, but the amount of carbon released was enhanced by the SA treatment during the SF period (Figure 5). For the R. soongorica litter, more carbon and nitrogen were released in the SA treatment during the SC period; in the SF period, the amount of carbon released differed among the three treatments, while the amount of nitrogen released did not. Across treatments, more carbon was released by S. passerina versus R. soongorica litter during the SC period, but this pattern was reversed during the SF period. During the SC period, less nitrogen was released by S. passerina than R. soongorica litter in the SA treatment.

3.4. Soil Nutrients and Enzyme Activity

Compared to CK, the soil organic carbon content associated with S. passerina litter was significantly lower in SR during the SC period, while the soil nitrogen content was higher in SA (Figure 6). For R. soongorica samples, the soil organic carbon content was significantly lower in SA than CK during the SF period, while the soil nitrogen content differed among the three treatments. During the SF period, the soil organic carbon content in SA was lower for S. passerina than R. soongorica samples, but the soil nitrogen content in both CK and SR was higher for S. passerina samples.

Snow addition (SA) generally enhanced the soil urease activity associated with both litter types and study periods, with the exception of R. soongorica samples during the SC period. The soil AKP activity was also higher in SA but only significantly so for S. passerina samples during the SC period. Meanwhile, snow removal (SR) reduced the soil urease activity for both litter types and study periods. The soil AKP activity was also reduced significantly by SR for both litter types during the SC period. Finally, soil CBH and POD activity were also enhanced in SA and diminished in SR for both litter types, with the exception of POD activity in R. soongorica samples during the SC period (Figure 6).

3.5. Factors Mediating Litter Mass Loss in Different Study Periods

In the SC period, the amount of C released and mass lost (ML) from both litter types was positively correlated with the soil moisture (SM), soil organic carbon (SOC), soil temperature (ST), and soil urease activity. In the SF period, the ML for both litter types was positively correlated with the CBH activity, POD activity, SOC, soil total nitrogen content (STN), ST, and urease activity. The amount of C released from S. passerina litter was positively correlated with SOC, ST, and STN, while the C released from R. soongorica litter was positively correlated with AKP activity, CBH activity, POD activity, SOC, STN, and urease activity (Figure 7).

Redundancy analysis (RDA) showed that soil characteristics and soil enzyme activity affected mass loss and nutrient release from both litter types (Figure 8). For R. soongorica litter, the first RDA axis explained 91.81% of the total variation and the second axis 4.79%; therefore, 96.6% of the variation in litter decomposition was explained by soil factors. Among the eight soil factors examined, the soil total nitrogen (STN) content was the most significant variable affecting both litter mass loss and nutrient release in R. soongorica (p = 0.002), explaining 88.6% of the total variance. Soil organic carbon (SOC) was the second important variable, explaining 5.6% (p = 0.002) of the total variance (Figure 8). For S. passerina litter, the first RDA axis explained 93.83% of the total variation, and the second axis accounted for 2.39%; soil factors, therefore, explained 96.22% of the variation in mass loss and nutrient release. Soil cellobiohydrolase (CBH) activity was the most significant variable affecting litter mass loss and nutrient release in S. passerina (p = 0.002), explaining 89.1% of the total variance. Soil organic carbon (SOC) was the second most important variable, explaining 4.4% of the total variation (p = 0.002) (Figure 8).

4. Discussion

4.1. Effects of Snowfall on Litter Mass Loss

Lower snow cover usually leads to a reduction in soil temperatures and heightened freeze–thaw cycles, which, in turn, affects litter decomposition [10]. In line with the first study hypothesis, greater snowfall (SA treatment) in the snow-covered (SC) period enhanced litter mass losses and the decomposition rate for both R. soongorica and S. passerina. This is consistent with most previous studies, which have found that a reduction in snow depth lowers the leaf litter decomposition rate [24,25,26]. This pattern may have several explanations. First, snow may act as an insulator, increasing soil temperatures and moisture levels [27] and thereby enhancing microorganismal activity in the winter [28]. Similarly, higher soil moisture levels may also promote litter decomposition in desert ecosystems. However, the micro-environmental changes associated with snow removal treatments are relatively large [29,30] and can lead to heavy decomposer losses [31], thus reducing decomposition (and overall litter mass loss). Second, soil moisture levels may be enhanced by melting snow cover, which is associated with an increase in the abundance and diversity of soil fauna; these changes can promote decomposition and litter mass loss [32,33]. Third, the leaf structure is destroyed by snow addition, leading to increases in plant surface area and improved substrate utilization for decomposers. For example, the degradation of cellulose and lignin was accelerated by snow addition [33], which resulted in stronger leaf litter decomposition [34]. Here, litter mass loss was positively correlated with soil temperature in the snow-covered period (Figure 6). This suggests that soil temperatures may shape litter decomposition in the winter in arid regions.

In the snow-free (SF) period, snowfall addition and reduction treatments had no effect on litter mass loss compared to the natural snowfall control, in contrast to the findings of previous studies [35]. A likely explanation for this discrepancy is that snowfall melts more rapidly in arid areas, influencing litter decomposition in the short term only. However, by the growing season, even though soil temperatures gradually increased, soil moisture did not differ among the three snowfall treatments. This led to similar levels of soil enzyme activity across treatments (Figure 6) and, therefore, similar litter mass losses in the SF period.

During the snow-covered (SC) period, snowfall addition enhanced the amount of carbon and nitrogen released from R. soongorica litter but had no effect on S. passerina litter. Snowfall removal did not affect carbon or nitrogen release from either litter type. These results only partly support the first study hypothesis and disagree with previous studies showing that reduced snow depth decreases the amount of C and N released for forest trees [36]. This might be due to the high initial nutrient content of R. soongorica litter; more nutrients may be released from nutrient-rich versus nutrient-poor leaf litter [37].

4.2. Effects of Snowfall on Leaf Litter Decomposition for Two Desert Species

In the snow-covered period, snowfall treatment effects on litter mass loss were the same for both study species, contrary to the second study hypothesis. However, in the snow-free period, treatment effects were greater for S. passerina than for R. soongorica, as predicted. The reason may be that snowfall treatments altered soil enzyme activity equally across study species in the SC period (Figure 6), with enzyme activity then mediating roughly even litter mass losses. Meanwhile, in the SF period, urease activity showed stronger treatment effects for S. passerina, potentially explaining the greater effects of snow addition on S. passerina litter. The correlation between litter mass loss and urease activity was also stronger for S. passerina (R² = 0.867) than R. soongorica (R² = 0.837) (Figure 7). Overall, these results suggest that snowfall addition promotes litter decomposition in the SC period, and this positive effect may be stronger for leaf litter with a higher C:N ratio [25].

4.3. Contrasting Snowfall Treatment Effects on Litter Mass Loss in the SC versus SF Period

The relative effects of snowfall treatments on litter mass loss differed between the SC and SF period. Litter mass losses were heightened in the SA treatment in the SC period, while mass losses were more prominent in the SR treatment in the SF period; this agrees with the findings of previous studies [28,38]. In arid regions, snow addition may increase soil temperatures in the SC period, promoting microbial activity and diversity, which ultimately enhances litter decomposition; this would explain SA treatment effects in the SC period [39]. In contrast, soil temperatures varied greatly within a day, and freeze–thaw cycles were more frequent in the SR treatment. This may have degraded the physical structure of the litter, thus promoting litter decomposition in the SF period [40].

5. Conclusions

Compared to the natural snowfall control, snowfall addition enhanced the litter decomposition rate and litter mass losses for both S. passerina and R. soongarica in the snow-covered period, while snowfall removal had the opposite effect. However, these effects differed between species during the snow-free period, with greater litter mass losses seen for R. soongorica versus S. passerina across snowfall treatments. Snowfall addition increased the amount of carbon and nitrogen released from S. passerina litter during the snow-free period and that released from R. soongorica litter during the snow-covered period. The soil total nitrogen content strongly shaped R. soongorica litter decomposition, while cellobiohydrolase activity drove S. passerina litter decomposition. Overall, snowfall changes significantly affected litter decomposition in both species, but these effects differed between study periods. As this study lasted one year, with single samples collected in the snow-covered and snow-free periods, these findings may reflect only the limited short-term effects of snowfall on litter decomposition. In future, more long-term experiments are needed, with multiple samples collected during the snow-covered period; such studies would improve our understanding of the mechanisms underlying snowfall effects on litter decomposition in arid desert regions.