*Article* **Water Uptake Pattern by Coniferous Forests in Two Habitats Linked to Precipitation Changes in Subtropical Monsoon Climate Region, China**

**Jianbo Jia 1,2,3, Yu Chen 1,2, Jia Lu 1,2 and Wende Yan 1,3,\***


**Abstract:** Variations in precipitation patterns under climate changes influence water availability, which has important implications for plants' water use and the sustainability of vegetation. However, the water uptake patterns of the main forest species under different temporal spatial conditions of water availability remain poorly understood, especially in areas of high temporal spatial heterogeneity, such as the subtropical monsoon climate region of China. We investigated the water uptake patterns and physiological factors of the most widespread and coniferous forest species, *Cunninghamia lanceolata* L. and *Pinus massoniana* L., in the early wet season with short drought (NP), high antecedent precipitation (HP), and low antecedent precipitation (LP), as well as in the early dry season (DP), in edaphic and rocky habitats. The results showed that the two species mainly absorbed soil water from shallow layers, even in the short drought period in the wet season and switched to deeper layers in the early dry season in both habitats. It was noted that the trees utilized deep layers water in edaphic habitats when the antecedent rainfall was high. The two species showed no significant differences in water uptake depth, but exhibited notably distinct leaf water potential behavior. *C. lanceolata* maintained less negative predawn and midday water potential, whereas *P. massoniana* showed higher diurnal water potential ranges. Moreover, the water potential of *P. massoniana* was negatively associated with the antecedent precipitation amount. These results indicate that for co-existing species in these communities, there is significant eco-physiological niche segregation but no eco-hydrological segregation. For tree species in two habitats, the water uptake depth was influenced by the available soil water but the physiological factors were unchanged, and were determined by the species' genes. Furthermore, during the long drought in the growing season, we observed probable divergent responses of *C. lanceolata* and *P. massoniana*, such as growth restriction for the former and hydraulic failure for the latter. However, when the precipitation was heavy and long, these natural species were able to increase the ecohydrological linkages between the ecosystem and the deep-layer system in this edaphic habitat.

**Keywords:** plant water source; habitat; stable isotope technology; leaf water potential; water use efficiency

#### **1. Introduction**

Increases in vegetation greenness have been reported around the world over the last three decades, manifested as the expansion of afforestation and reforestation [1–3]. However, forests may be vulnerable to degradation due to global climate changes with new precipitation patterns [4–6]. Changes in the characteristics of precipitation may result in changes in water availability, which have implications for plants' water uses in ecosystems [7,8]. The variations in plants' water use responses to precipitation and water

**Citation:** Jia, J.; Chen, Y.; Lu, J.; Yan, W. Water Uptake Pattern by Coniferous Forests in Two Habitats Linked to Precipitation Changes in Subtropical Monsoon Climate Region, China. *Forests* **2022**, *13*, 708. https://doi.org/10.3390/f13050708

Academic Editor: Steven McNulty

Received: 1 April 2022 Accepted: 29 April 2022 Published: 30 April 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

availability plays important role in the sustainability of the restored vegetation and the promotion of the water cycle in critical zones [9–11].

The temporal–spatial heterogeneity of precipitation and water availability affect plant water use strategies [12,13]. At the point scale, the water source variability along the soil profile is one of the most important factors for water uptake by plants [14]. At the surface scale, the aquifer storage is distinct in different habitats, such as deep soil habitats [15], outcrop habitats [16], and soil with rock fragments habitats [17], which is related to the soil properties and plant water consumption. At different stages of the same season, the plant water uptake depth may also differ with changes in rhizosphere water availability [15,18]. Meanwhile, the amount of precipitation may be a critical factor affecting the water sources of trees. The plant water uptake can be identified by contrasting the δD and δ18O of xylem water and all the potential water sources. Previous studies have shown that tree species may switch their water sources from shallow layers in the wet season with sufficient precipitation to stable layers in the dry season using stable isotope techniques [19–21] Liu [22] found that following rainfall events, *Platycladus orientalis* L. trees with a dense and shallow fine root system absorbed more water from the soil surface layers and precipitation. Other plants, however, mostly take up water from deep and stable layers regardless of seasonal changes or precipitation events in the semi-arid regions [16,23]. In contrast, in subtropical regions, evergreen species use shallow soil water with a drought-avoidance strategy even under seasonal drought conditions [24].

The divergent response of plant water uptake to changes in precipitation and water availability has been related to physiological characteristics. It has been suggested that the predawn and midday leaf water potential can be used to describe the daily patterns of plantwater relations, coupling water among the root zone, the plant itself and the atmosphere [25]. Previous studies have shown that plants relying on shallower water sources exhibited a larger diurnal range of leaf water potential, and on the contrary, narrower diurnal ranges are usually linked with deep and stable water sources [26–28]. Moreover, the plant water efficiency (WUE) has attracted attention as a means of reflecting plant water use characteristics, together with plant water uptake [29,30]. Nie [31] explored leaf WUE based on δ13C values and found that the high WUE corresponded with the use of deep water sources, indicating more conservative water-use strategies in a subtropical monsoon climate region. The plant water uptake pattern was found to be influenced by water availability and physiological traits in different ecosystems [32]. However, the relationship between these two factors affecting plant water uptake is unclear, especially in complex and fragile forest ecosystems, which limits the understanding of restored vegetation adaptability and rock-soil-water-plant-atmosphere interactions in critical zones.

Subtropical China, which is characterized by a monsoon climate, is an ecologically sensitive area that is affected by global changes [33]. The precipitation in this region is abundant and the alternation of dry and wet is obvious. The change in precipitation patterns has led to a reduction in the available water in the ecosystem, and the risk of drought stress and drought death has significantly increased [4]. The distribution of the plantations is a clustered distribution with heterogeneous habitats (such as thin soil habitats with rock fragments, and outcrop habitats with soil fragments) [34]. Different rock and soil structures could influence hydrological processes and the amount of soil water available. Plant water use strategies in different habitats are critically important for the evaluation of vegetation adaptation. A number of previous studies have primarily focused on the water sources of different types of plantations or natural vegetation in one specific habitat [16,35]. Few have paid attention to differences in plant water uptake patterns in different habitats, which has limited our understanding of plant water adaptation and the evaluation of sustainable vegetation restoration. With changes in the global precipitation pattern, shortterm drought and rainstorms have become more frequent, especially in the wet/growing season. However, it is unclear how the water uptake of plants in the different habitats responds to these precipitation changes.

Based on the above analysis, we applied stable isotope techniques (δD and δ18O) and measurements of leaf water potential to determine coniferous-leaved forest water uptake patterns in two habitats (edaphic and rocky habitats) with different antecedent precipitation during the growing season in the subtropical monsoon climate region of China. The main objectives of this study were: (i) to investigate the responses of belowground water use patterns of *Cunninghamia lanceolata* L. and *Pinus massoniana* L. to the temporal– spatial heterogeneity of water availability, as shown in conditions of different antecedent precipitation levels and edaphic and rocky habitats; and (ii) to understand the aboveground physiological responses to varied water availability of two species, analyzed by examining the variations in leaf water potential behavior and water use efficiency. Our first hypothesis was that soil water availability could have an effect on the plant water uptake depth, and that the two species may show similar water sources, and the second was that the plants' physiological factors may vary with the changes in water availability and species types.

#### **2. Materials and Methods**

#### *2.1. Study Sites*

This study was conducted at the Hunan Lutou forest ecosystem observation and research Station (28◦31 7–28◦38 N, 113◦51 52–113◦58 24 E) (Figure 1). The region has a subtropical mountainous monsoon climate, with a mean annual precipitation of 1450.8 mm and an annual temperature of 18.5 ◦C. The wet season, which receives more than 60% of the annual rainfall, lasts from late April to late September, and the dry season extends from December to February [28]. The growing season spans from April to October. The study area is dominated by *Cunninghamia lanceolata* and *Pinus massoniana* secondary forests. The understory contains species such as *Fortunearia sinensis* Rehd, *Ilex cornuta* and *Asparagus cochinchinensis*, and the forest coverage rate is more than 90%. Soil in the study area is predominantly red soil, having a general soil layer that is 80–100 cm thick. The other part of the slope has a high exposed rock ratio whereas the soil occurs discontinuously, only in rock gaps. Thus, the habitats were variable, with the different outcrop ratios, such as an edaphic habitat with a low outcrop ratio, a continuous broken rock habitat with patches of soil, an isolated outcrop habitat, and so on. Springs sometimes appear at the bottom of hillslopes during the rainy season or after rains in the drought season.

**Figure 1.** Map of the study and the field area. (**a**): The location of Lutou forest ecosystem observation and research Station in Hunan province; (**b**): The sampling site location in the study area; (**c**) photographs of the two habitats.

According to the distribution of these typical habitats, one habitat consisting of thick soil with rock fragments habitat (the "edaphic habitat" for short) and the another habitat consisting of continuous stone outcrops with soil fragments (the "rocky habitat" for short) were chosen at the foot of the Northwest-facing hillslope in two 20 m × 20 m sample plots. The two habitats were 250 m apart whereas the elevation difference was about five meters. In the edaphic habitat, the soil was relatively thick (about 90 cm deep), horizontally interrupted by small outcrops, and vertically interrupted by small rocks. Along the soil profile, the upper layer soil (0–30 cm) was well-drained whereas the lower layers (30–70 cm) were sticky with a low saturated hydraulic conductivity (*K*s). Underneath the soil was a high-weathered dolomite bedrock zone (70–90 cm). The outcrop ratio was about 20% in this habitat. In the rocky habitat, the outcrop ratio was more than 80%, and the range of height from the top of the outcrop to the soil in the rock gaps was from 0.3 m to 3 m. The soil was inlaid in the rock in a spotty pattern and was discontinuous (average 30 cm deep on average). Similarly, a high-weathered bedrock zone was present under the soil. The vegetation was sparse in this habitat. There was an intermittent spring outflow near the two habitats at the bottom of the hillslope.

For the subsequent analysis and comparison, the plant water sources were divided into shallow (0–30 cm), middle (30–70 cm in the edaphic habitat and 30–50 cm in the rocky habitat), and deep (70–90 cm in the edaphic habitat and 50–70 cm in the rocky habitat) layers and spring according to the soil texture and fluctuations and patterns of isotopic ratios in the soil water, VWC, and the impact of the rainfall pulse. (1) Shallow soil layer: The variability of soil water isotopic compositions and VWC in this layer was larger, and it was vulnerable to rainfall pulses and evaporation with seasons. (2) Middle soil layer: The variability of soil water isotopic compositions and SWC in this layer was lower than that of the 0–30 cm soil layer. The impacts of rainfall pulses and evaporation were moderate. Both the clay content and soil bulk density were higher than the shallow layers. (3) Deep soil layer: This layer was high-weathered bedrock with high leakage and low water holding capacity in the rocky habitat and high water moisture in the edaphic habitat, respectively.

#### *2.2. Plant and Soil Sampling*

In order to explore the relationship between plant physiological traits and water uptake patterns for adapting to precipitation change, plant and soil sampling were conducted simultaneously at the two habitats bimonthly on 12 June (wet season with high antecedent precipitation, HP), 5 August (wet season with low antecedent precipitation, LP), and 18 October (early dry season, DP) 2020. We also sampled on 18 May in the early wet season with a 20 day drought (no rain, NP). Two coniferous species, *Cunninghamia lanceolata* (DBH of from 5 to 11cm, average DBH was 7.9 cm) and *Pinus massoniana* (DBH of from 6 to 12cm, average DBH was 8.5cm) in each of the habitats were selected for the study. We selected four individuals per species for analysis, and the DBH of sampled trees were used to represent the average DBH in the stands. The leaf and plant xylem samples from every selected plant were collected in each habitat. Every selected plant was collected in each stand-age tree per month. The fully sun-exposed, mature and healthy leaves in the upper canopy from each selected plant were collected in different directions on each sampling date. The leaves were mixed and packed into craft paper bags and brought them back to the laboratory for the measurement of the plant leaves' δ13C levels. Shoots ranging from 0.3 to 0.5 cm in diameter and 3 to 5cm in length were collected at mid-day from stems that were more than 2 years old [28]; the outer bark and phloem of the shoots were removed to obtain the xylem sample.

Soil samples were obtained in two habitats from six depth intervals (0–10, 10–20, 20–30, 30–50, 50–70, 70–90 cm) with an auger (sampling only at 70 cm deep in the rocky habitat) and five replicates were collected at each layer. Among them, the high-weathered bedrock samples were collected between 70–90 cm in the edaphic habitat and 50–70 cm in the rocky habitat. A subsample of the soil samples was stored at −20 ◦C for isotopic analysis, whereas the remainder of the samples were sealed for the measurement of gravimetric soil water content, obtained by oven drying for one day. The volumetric water content (VWC) was converted according to gravimetric water content and bulk density (Table 1) of each layer.


**Table 1.** The soil bulk density of two the habitats.

#### *2.3. Precipitation and Spring Sampling*

Water samples were routinely collected for each rain event above 5 mm from May 2020 to December 2020. The isotopic values of precipitation were not collected from January to April due to the COVID-19 pandemic. The collection equipment was designed based on the new device for monthly rainfall sampling developed for the Global Network of Isotopes in Precipitation [36]. The rainwater samples were stored in cap vials, wrapped in parafilm, and stored in a freezer until the analysis of stable isotopes. Data on temporal distribution of rainfall data and other meteorological data were collected at a meteorological station located in the middle of the same small catchment. Spring water discharged from1 June to 29 November, but were cut off between 25 July to 29 August. The spring was sampled regularly during the outflow period. Both rainwater and spring water were stored in cap vials, wrapped in parafilm, and frozen until stable isotope analysis.

#### *2.4. Isotopic Analyses*

The water was extracted from the xylem and soil using an automatic cryogenic vacuum distillation water extraction system (LI-2100, LICA, Beijing, China) [37,38]. The δD and δ18O in the xylem and soil water samples were measured with liquid water isotope ratio infrared spectroscopy (IRIS, DLT-100, Los Gatos Research, Mountain View, CA, USA) at the Key Laboratory for Agro-Ecological Processes in Subtropical Region, Chinese Academy of Sciences. The δ13C level in the plant leaves were analyzed using an isotope ratio mass spectrometer (IRMS, MAT253, Thermo Fisher Scientific, Bremen, Germany).

The isotope composition is reported in δ notation relative to V-SMOW as

$$\text{\\$X = (R\_{\text{sample}}/R\_{\text{standard}} - 1) \times 1000} \tag{1}$$

where X represents D, 18O, or 13C. Rsample and Rstandard are the ratios of D/H, 18O/16O, or 13C/12C ratio of a measured sample and a standard sample, respectively. The standard deviation for repeat measurements was ±1‰ for <sup>δ</sup>D, ±0.2‰ for <sup>δ</sup>18O, and ±0.15‰ for δ13C.

Extracting water from the plant xylem using cryogenic vacuum distillation can result in the mixing of organic materials (e.g., methanol and ethanol), which may affect the spectroscopy and lead to erroneous stable isotope values when analyzing them with IRIS [39,40]. We have corrected the isotopic values of the xylem according to Liu [28].

#### *2.5. Leaf Water Potential*

Predawn and midday water potentials (Ψpd and Ψ md, respectively) of leaves were measured in the wet seasons (simultaneously with isotope sampling) with a pressure chamber (PMS Instruments Co., Corvallis, OR, USA). Samples (*n* = 5 per species) were collected from branches that were fully exposed to the sun, at places where branches were 2/3 of the way up of the canopy, at least 2 m above ground. The measurements were performed between 4:00 to 6:00 h for predawn water potential and between 12:00 and 14:00 h for midday water potential on the same day.

#### *2.6. Data Analysis*

Plant water source partitioning was determined by means of the Bayesian mixing model MixSIAR (version 3.1.7) [41]. The raw isotopic ratios of the xylem water were input into MixSIAR as the mixture data. The averages and standard deviations of the soil water isotopes in the different soil layers were the source data. The discrimination was set to zero for both δD and δ18O because there is generally no isotopic discrimination of water during plant water uptake by roots [42].

Independent-samples *t*-tests and One-way ANOVA were used to detect the differences in plant water sources and water potential among the species, habitats and their seasonal differences. Post hoc comparisons were based on Tukey's HSD. Moreover, Pearson correlation was used to conduct the correlation analysis, and the figures were plotted with Origin 9.0 software (Origin, Origin Lab, Farmington, ME, USA).

#### **3. Results**

#### *3.1. Isotopic Compositions of Precipitation, Soil Water, and Springs*

The total precipitation was approximately 2121 mm in 2020 and the distribution of rainfall was temporally uneven, with 79.32% of the rainfall occurring during the wet season (Figure 2). It was noted that there are two extreme precipitation events occurred—on 7 in September (282.2 mm) and on 7 June (115.2 mm). Except for the NP sampling with a 20 day drought, the accumulated precipitation amounts ten days before the last three samplings were 283.6 mm, 49.4 mm, and 55.4 mm, respectively.

**Figure 2.** Variations in precipitation, mean air temperature, and isotopic values (δD, δ18O) in precipitation at a daily timescale in 2020. Arrows indicate sampling dates. (The isotopic values of precipitation were not collected from January to April due to the COVID-19 pandemic impacting).

The isotopic compositions of the precipitation showed a large fluctuation during the study period (Figure 2). The mean <sup>δ</sup>D of the precipitation was −48.69 ‰, the mean <sup>δ</sup>18O of the precipitation was −7.88 ‰. The relatively depleted isotopic values of precipitation occurred when it rained continuously for a long time, with high precipitation. The δD level obtained for ten days of precipitation before the three samplings were ranged from −23.55‰ to −57.52‰, −34.54‰ to −68.36‰, −40.76‰ to −51.02‰, respectively. The <sup>δ</sup>18O of precipitation before three samplings were ranging from −5.27‰ to −8.15 ‰, −7.6 ‰ to −9.65 ‰, −6.54‰ to −7.4‰, respectively.

The δD and δ18O values of soil water in the different habitats varied with soil depth and season (Figures 3 and 4). In the edaphic habitat, the average δD value of the soil water was −45.56‰ ± 16.05 ‰ (mean ± S.D.), and the average <sup>δ</sup>18O value was −6.55‰ ± 1.73‰. The average <sup>δ</sup>D and <sup>δ</sup>18O values of soil water in the rocky habitat were −44.6‰ ± 16.58‰ and −6.7‰ ± 1.96 ‰, respectively. There were no significant differences (*p* = 0.84 for δD, *p* = 0.79 for δ18O) in the soils' isotopic compositions in the different habitats. In NP, the soil water isotopes were observed to be depleted with soil depth (Figures 3a and 4a). In HP, the δD and δ18O values of water along the soil profile were consistent with recent rainfall values (Figures 3b and 4b). In the two late two samplings, the soil water isotope composition converged at the top and bottom layers, which were similar to recent rainfall values (Figure 3c,d, and Figure 4c,d). The middle-layer soil water showed more enriched values in LP and depleted isotopic values in DP and exhibited less variation with soil depth. There were no significant differences obtained for soil water (*p* = 1.28 for δD, *p* = 0.93 for δ18O) in different sample layers (i.e., 0–10 cm, 10–30 cm, 30–50 cm, 50–70 cm, 70-90 cm). However, when merging sample layers into shallow, middle and deep layers (see Methods), the soil water isotope was significant different in the two habitats (*p* < 0.05).

**Figure 3.** Variation in mean (±S.D.) δD of soil water with the soil profile, precipitation, spring, and xylem water during the wet season (**a**) May sampling; (**b**) June sampling; (**c**) August sampling; (**d**) October sampling.

The isotopic composition of springs changed across the sampling time. The isotopic values were less negative in HP than in DP. The δD and δ18O values of xylem water were less negative in NP and became more negative with the seasonal changes. There were no significant differences in isotopic composition between species types and habitats (*p* > 0.05), except in June-HP, when the xylem water isotope was more negative in the rocky habitat than that in the edephic habitat (*p* < 0.05).

**Figure 4.** Variation in mean (±S.D.) <sup>δ</sup>18O of soil water with the soil profile, precipitation, spring, and xylem water during the wet season (**a**) May sampling; (**b**) June sampling; (**c**) August sampling; (**d**) October sampling.

#### *3.2. Variations in Soil Water Content, Water Uptake Patterns, and Their Linkage with Precipitation*

The VWCs of the two habitats displayed clear vertical and seasonal variations (Figure 5). The average VWCs were 43.42% ± 7.68% in the edaphic habitat and 38.24% ± 8.42% in the rocky habitat, with no significant differences (*p* = 0.07) during the study periods. However, the VWCs of shallow soil layers in the two habitats differed significantly (*p* < 0.001). In NP, the VWC of the shallow layer was the lowest in the two habitats and the soil moisture increased with depth (Figure 5a). Furthermore, the VWC exhibited a slightly increasing tendency along the soil profile in the edaphic habitat but a decreasing tendency in the rocky habitat with the seasonal changes. It was noted that the soil moisture in the edaphic habitat was significantly higher than that in the rocky habitat in LP, especially in the middle layers, which may be related to the different soil texture and plant transpiration characteristic (Figure 5c).

The two tree species mainly took up soil moisture throughout the wet season in two habitats, and the proportions of water sources used by the two species exhibited no significant differences (*p* > 0.05) (Figure 6). However, the plant water uptake depth differed between habitats and across seasons. In HP, trees in the rocky habitat absorbed more than 67.14% of their water from shallow soil layers, whereas the mean water uptake ratio of the two tree species in the edaphic habitat were 64.45% for the middle and deep soil layers. In DP, the *C. lanceolata* and *P. massoniana* in edaphic habitat obtained more than 74.82% of their water from the shallow and deep soil layers. On the other hand, in the rocky habitat, the two species mainly extracted soil water from shallow and middle layers (82.13%). In NP and LP, both *C. lanceolata* and *P. massoniana* in the two habitats utilized the largest proportion of shallow soil water (64.97%, 0–30cm).

**Figure 6.** Variation in mean (±S.D.) water source proportions for *C. lanceolata and P. massoniana* during the wet season.

The responses of the proportion of plant water sources used in each soil layers to the amount of precipitation ten days before sampling were distinct in the two habitats (Figure 7). In the edaphic habitat, tree species absorbed less water from shallow layers and absorbed more deep soil water with the increases in precipitation (Figure 7a,c). On the other hand, the trees maintained a high water uptake from shallow layers in the rocky habitat regardless of precipitation variations (Figure 7d). Meanwhile, there were significant negative linear relationships between the water source proportions of the middle and deep soil layers and precipitation (Figure 7e,f).

#### *3.3. Variation in Leaf Water Potential and Its Linkage with Precipitation*

The *ψ*pd of the two species was found to be less negative (>−1MPa) in the two habitats, indicating no severe water stress during the study period, whereas the *ψ*md was lower than *ψ*pd, away from the 1:1 line, and exhibited profoundly seasonal variations (*p* < 0.01) (Figure 8). Both of the two species showed lower *ψ*md values in NP and DP than that in HP and LP (*p* < 0.05). Furthermore, *P. massoniana* showed significantly more negative *ψ*md values than *C. lanceolata,* especially in NP and DP (*p* < 0.05). However, there were no significant differences in *ψ*pd and *ψ*md between the edaphic and rocky habitats for the two species (*p* > 0.05).

**Figure 7.** Relationships between water source proportions for each soil layers (mean ± S.D.) and the precipitation amount ten days before sampling. P is Pearson correlation, *R*<sup>2</sup> represents the fitting degree of the relationship between the water source proportion and the precipitation amount; *p* is the *p*-value of the fitting ((**a**–**c**) plant water sources from the shallow, middle, and deep layers in the edaphic habitat, respectively; (**d**–**f**) plant water sources from the shallow, middle, and deep layers in the rocky habitat, respectively). The black dots represented *C. lanceolata*, and the red dots represent *P. massoniana*.

**Figure 8.** Plot of mean (±S.D.) predawn water potential and midday water potential of two tree species in the edaphic and rocky habitats. The black line is the 1:1 line of predawn water potential vs. midday water potential. The light to dark colors for each species represent the sampling dates as May-NP, June-HP, August-LP, and October-DP. The light orange shadow represents the cluster of trees close to the 1:1 line in HP and LP. The light blue shadow represents the cluster of trees away from the 1:1 line in NP and DP.

The diurnal ranges of water potential (Δ*ψ*) exhibited significant variation in different species with seasonal changes (*p* < 0.01) (Figure 9). *C. lanceolata* showed significantly lower Δ*ψ* values than *P. massoniana* (*p* < 0.001). The Δ*ψ*max value was the highest in NP for *P. massoniana* (−1.84 ± 0.19 MPa) and in DP for *C. lanceolata* (−0.45 ± 0.34 MPa). Both of the two tree species displayed the minimum Δ*ψ* (−0.48 ± 0.11 MPa and 0.09 ± 0.06 MPa, respectively) in HP and LP. Both of the two species showed significantly higher diurnal ranges of water potential in the edaphic habitat than those in the rocky habitat (*p* < 0.001) during the sampling period, except for *P. massoniana* in LP and DP. Furthermore, there was no significant correlation between the Δ*ψ* values and water uptake depth for *C. lanceolata* or *P. massoniana* in the two habitats.

**Figure 9.** Variation in mean (±S.D.) diurnal ranges of water potential for *C. lanceolata and P. massoniana* during the wet season.

The high vapor pressure deficit and solar radiation showed a strong atmospheric evaporation force (AEF) (Table 2) that influenced transpiration and the diurnal changes in water potential. The meteorological factors during the sample period showed no significant variation (*p* > 0.05), except for DP with lower AEF. On the other hand the Δ*ψ* values exhibited seasonal changes and showed the highest values in NP and DP for the two species. These changes may be affected by the available of soil water. The responses of the Δ*ψ* to the precipitation amount ten days before sampling were different in the two tree species (Figure 10). The Δ*ψ* values of *C. lanceolata* did not increase with the change in conditions from no rain to high rainfall in the two habitats. However, the Δ*ψ* values of *P. massoniana* showed lower values with the precipitation increases in the edaphic and rocky habitats. Moreover, the plant water uptake depth was not correlated with the diurnal range of water potential (Table 3).


**Table 2.** The variations in meteorological factors during the sampling dates.

**Figure 10.** Relationships between the diurnal ranges of water potential (mean ± S.D.) and the precipitation amount ten days before sampling. P is Pearson correlation, *R*<sup>2</sup> represents the fitting degree of the relationship between the diurnal ranges of water potential and the precipitation amount; *p* is the *p*-value of the fitting ((**a**) *C. lanceolata* in the edaphic habitat; (**b**) *P. massoniana* in the edaphic habitat; (**c**) *C. lanceolata* in the rocky habitat; (**d**) *P. massoniana* in the rocky habitat).



#### *3.4. Variation in Leaf Water Use Efficiency*

The δ13C values of the two species were significantly more negative in the middle wet season and early dry season than that in May-NP (*p* < 0.05) (Table 4). The leaf water use efficiency was higher in the drought stage in the wet season. There were no significant differences (*p* > 0.05) in δ13C values between the two habitats. Furthermore, with the exception of in DP in the edaphic habitat and in LP in the rocky habitat, *C. lanceolata* and *P. massoniana* showed no significant differences in δ13C.


**Table 4.** Comparisons of δ13C values of *C. lanceolata* and *P. massoniana* in edaphic and rocky habitats.

Note: Capital letters represent significant differences of the same tree species among different sampling dates at the 0.05 level; lowercase letters represent significant differences between *C. lanceolata and P. massoniana* in the same habitat at the 0.05 level.

#### **4. Discussion**

#### *4.1. Water Uptake of Tree Species in Two Habitats*

The variation of plant water uptake depth in the two habitats was consistent, except in June-HP. These two species, growing at the foot of the slope, mainly absorbed soil water from shallow layers in the early and middle wet season, and switched to deeper layers in the late wet season. This water uptake pattern has also been observed in other natural species and plantations in the similar study areas [16,28]. However, it was noted that the plants utilized shallow soil water rather than deep water (no springs flowing) in the early wet season with a 20 day drought, which was inconsistent with other studies in this climate region [16,22]. Although the mean soil moisture was lower compared to other samplings, the VWC was still higher than that observed in semiarid climate regions in the wet season [15,43]. Meanwhile, with a relatively lower wilting coefficient and high spatial heterogeneity [44], the shallow layers could also provide enough available water for plants. Previous studies showed that the plant species adjusted their physiological factors, such as water potential behavior, water use efficiency, in response to the environment changes [25,45,46]. In our study, *C. lanceolata* and *P. massoniana* exhibited the highest leaf Δ*ψ* and δ13C values in NP, indicating that they tried their best to absorbed enough shallow soil water with lower midday leaf water potential to balance carbon-water relations in tandem with high leaf-level intrinsic water use efficiency (iWUE). Moreover, this water uptake pattern is an adaptation, enabling plants to save more energy for growth in the early wet season. Both *C. lanceolata* and *P. massoniana* grow quickly, showing high energy consumption in May, as well as in the early growing season. Although the deep soil layer has a higher VWC, the energy required to take up water from the deep layer is greater than that of the upper layers [47,48]. Thus, the trees extracted shallow soil water to avoid excessive energy consumption through physiological adjustments [49–51]. In the middle and late wet season, plants water uptake depth shifted from shallow to deeper layers. Soil water availability may be the main reason for this water uptake pattern [52,53].

When the antecedent precipitation was much higher in the middle wet season, the plants still absorbed water from shallow layers in the rocky habitat, but in the edaphic habitat they switched to deep layers of soil water. Water availability is the most important factor influencing the plants water uptake depth [26,54]. Soil structure, such as soil texture, bulk density, affected water holding capacity, and migration, along with soil profiles, thus regulated plant water use [28,55]. The bulk density in the rocky habitat was lower than that in the edaphic habitat, promoting the high water holding capacity, whereas in the thin deep layers with large cracks and crevices in the rocky habitat, moisture leaked into the layer, flowing through the springs. In the thick deep layer with fine cracks in the edaphic habitat, the amount of stored water was higher than that in the shallow layer after large and continuous precipitation. Therefore, discrepancies in soil properties are the main reasons underlying the different soil water availability along the profiles in the two habitats. Furthermore, the low diurnal ranges of water potential of *C. lanceolata* and *P. massoniana* also demonstrated that they were both had sufficient water supplies in the two habitats.

#### *4.2. Water Uptake Pattern and Physiological Factors Change in the Different Tree Species*

The two coexisting plants—both in the edaphic and the rocky habitat—exhibited no significant differences in water uptake patterns, indicating that they had the same ecohydrological niche and no water source segregation. This result was consistent with a previous study in a similar climate region, which showed that the six mixed plantations had similar water sources, using the 0–30 cm soil layers in the wet season [28]. Studies in other regions also showed that coexisting species usually exhibited water competition in mixed stand [56,57]. Nie [58] investigated three communities on adjacent rocky hill slopes, and found that different species within each community all exhibited the use of a similar water source. Du [59] studied three karst climate forest communities of a typical hill, and obtained the same results. The similar root distribution of *C. lanceolata* and *P. massoniana* may be the main reason that they exhibited the same water uptake pattern [60,61]. Hence, the interspecific difference in community was relatively low in the subtropical monsoon climate region. However, as per the above analysis, the water uptake pattern was different between the edaphic and rocky habitats for the same species. This suggested that the habitats may have more of an influence on plant water use than the interspecific differences in the community, especially when the antecedent precipitation is high.

Although the water uptake depth was similar for the two species, the two species had different physiological responses to the water uptake. In our study, *C. lanceolata* maintained small diurnal ranges of water potential, high leaf δ13C values, and a large amount of branching from the base of the trunk, whereas *P. massoniana* showed the inverse characteristics. Meanwhile, the Δ*ψ* values of *P. massoniana* in the two habitats were negatively associated with antecedent precipitation, but a significant relationship was not observed in *C. lanceolata*. Wang [15] found similar results for two species in a mixed plantation in the Loess Plateau. Moreno-Gutiérrez [50] observed the existence of species-specific ecophysiological niche segregation in dryland plant communities. A possible explanation was that the inter-specific competition in the same habitat caused each tree species to establish different hydrological niches for water uptake [48,62]. Unlike the previous findings, in our study, there was significant eco-physiological niche segregation but no ecohydrological segregation for the two species in the same habitat. The plant water uptake depth was not correlated with the diurnal range of water potential. In other words, the aboveground physiological parameters showed significantly differences between two species, whereas the belowground water uptake was consistent among the two species. This discrepancy may be attributed to sufficient precipitation and soil water availability for ecohydrological non-segregation [63] and interspecific differences in terms of eco-physiological segregation [54].

#### *4.3. Implications for Plant Water Adaptation under Precipitation Changes*

With the increasing temperatures, precipitation patterns change seasonally and become more variable [8], which could lead to an increase in either the severity of drought or extreme precipitation, especially in the growing season [64–66]. When drought or extreme precipitation occurs, soil water availability may influence the plants' water use strategies.

In our study, plants absorbed soil water from shallow layers by increasing the diurnal ranges of water potential and water use efficiency in the early wet season with a 20 day drought. The tree species sought a balance between water uptake and growth through their relatively high water use efficiency [67]. However, if the drought was prolonged, soil moisture would decline and fail to supply water for plants. Ding [26] conducted a 135 day rainfall exclusion experiment in a catchment, and found two adverse responses, according to different physiological characteristics, to the severe water limitation: canopy defoliation and/or mortality and survival. In our study, *P. massoniana*, as the species exhibiting profligate water use exhibited larger Δ*ψ* and lower *ψ*md values for absorbing water sources [26]. Once the *ψ*md values beconmes lower than the hydraulic trait values, the species may suffer from the risk of hydraulic failure, such as xylem cavitation and leaf turgor loss [56,68]. On the contrary, *C. lanceolata* displayed stable Δ*ψ* values in the

sampling period, indicating the rigorous stomatal control [69]. The tree growth rate of *C. lanceolata* may slow due to the reduction in shallow soil water sources and the advanced stomatal closure.

Except for the drought in the growing season, the frequency of rainstorms and extreme precipitation also has also been increasing recently [70,71]. Plants are the main conduit for returning terrestrial water to the atmosphere, thereby exerting a strong effect on hydrologic fluxes of the terrestrial-atmospheric system [63]. In our study, the plants that mainly utilized for deep layer soil water in the edaphic habitat and the Δ*ψ* values of *P. massoniana* were lower when the precipitation was extremely high. These results illustrated that the tree species could adjust their water use strategies and increase the eco-hydrological linkages between the ecosystem and the deep-layer system [59].

#### **5. Conclusions**

In this study, the stable isotope technique and a pressure chamber were applied to detect the seasonal water uptake patterns of two coniferous species in edaphic and rocky habitats in a subtropical monsoon climate region. The results showed that the two species mainly absorbed soil water from shallow layers, even in the short drought period in the wet season and switched to deeper layers in the early dry season. It was noted that the trees utilized deep-layers water in edaphic habitats when the antecedent rainfall was high. The two species showed no significant differences in water uptake depth, but notable differences in their leaf water potential behaviors. *C. lanceolata* displayed narrow and stable Δ*ψ* values whereas the Δ*ψ* values of *P. massoniana* were negatively associated with antecedent precipitation. Thus, for co-existing species in communities, there was significant eco-physiological niche segregation but no eco-hydrological segregation. For the tree species in the two habitats, the water uptake depth was influenced by the soil water availability, but the physiological factors were unchanged, determined by the species genetics. Furthermore, during a long drought in the growing season, *C. lanceolata* and *P. massoniana* probably show divergent responses, such as growth restriction and hydraulic failure. However, when the precipitation is heavy and long, these species could increase the ecohydrological linkages between the ecosystem and the deep-layer system in the edaphic habitat.

**Author Contributions:** J.J. and Y.C. conducted field experiment, performed data analysis, and wrote the draft manuscript. J.L. and W.Y. conceived the study, designed the experiment. All authors contributed to discussion and interpretation of resulting data. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the key research and development program in Hunan province (2020NK2022), the National Natural Science Foundation of China (41807162) and the Hunan Province Natural Science Foundation (2019JJ50994).

**Data Availability Statement:** All relevant data are within the manuscript.

**Conflicts of Interest:** We declare that we have no financial and personal relationships with other people or organizations that can inappropriately influence our work, there is no professional or other personal interest of any nature or kind in any product, service and/or company that could be construed as influencing the position presented in, or the review of, the manuscript entitled "Water uptake pattern by coniferous forests in two habitats linked to precipitation changes in subtropical monsoon climate region, China".

#### **References**


**Roman Gorbunov \*, Vladimir Tabunshchik, Tatiana Gorbunova and Mariia Safonova**

A.O. Kovalevsky Institute of Biology of the Southern Seas of RAS, 299011 Sevastopol, Russia

**\*** Correspondence: karadag\_station@mail.ru

**Abstract:** This article discusses the processes of moisture intake, redistribution, and consumption within the downy oak forest community, along with their interannual and interseasonal water balance dynamics. The study of the water balance components was conducted using a combination of field research methods and geoinformation modeling on the territory of the Karadag landscape and ecological station of the Karadag Nature Reserve for the period from 2010 to 2020. The study of the water balance of downy oak forests located at the furthest extent of their range represents an important problem, whose solution will further scientific understanding by uncovering individual patterns of the internal organization of such systems. The indicators having the most tangible impact on the water balance are the amount of precipitation and evapotranspiration. The average annual precipitation on the territory of Karadag for the analyzed period was 448 mm; in recent years, a decrease in the amount of precipitation has been recorded. The evapotranspiration values within the downy oak forests approximately coincide with the values of this indicator in the Mediterranean region to average 450 mm per year. The influence of stemflow and relief features on the redistribution of moisture within the landscape is described. The analyzed water balance components' dynamics form conditions conducive to the displacement of steppe communities by forest species.

**Keywords:** water balance components; downy oak forests; landscape; Southeastern Crimea; precipitation; evapotranspiration; runoff

#### **1. Introduction**

The study of the Mediterranean climate and the features of its radiation, heat, and water balance are highly topical issues [1–5], since this region is located on the border of an arid climate and a temperate and rainy climate [6]. The theoretical foundations for studying the water balance of the territory of the Crimean Peninsula are laid out in the works of [7–13]. An example of evaluating the regional and local water balance characteristics in Crimea can be found in the works of [9,14–17]. At the same time, very little attention has been paid to the water balance of certain types of plant communities growing at the local level on the territory of the Crimean Peninsula.

Despite the Crimean Peninsula being located far beyond the Mediterranean Basin, the southern coast of Crimea, along with other regions of the Caucasus, is defined as an exclave of Mediterranean vegetation [18–20].

On the territory of the Crimean Peninsula, subtropical forest landscapes are mainly represented by downy oak and juniper formations with an admixture of Christ's thorn, smoke tree, and Oriental hornbeam. These are represented from the coast to an altitude of 300–350 m [9], where they are either replaced by typical downy oak forests and their derivatives with an admixture of hornbeam, higher rock-oak forests with an admixture of hornbeam and ash forests, or rise to the edge of the steep southern limestone slope of the yayla (upland pasture). These communities are typically represented by shiblyak thickets, comprising a dense low-growing closed forest area, or, conversely, sparse woodlands. The Crimean shiblyak and phrygana biomes are often identified with the Mediterranean

**Citation:** Gorbunov, R.; Tabunshchik, V.; Gorbunova, T.; Safonova, M. Water Balance Components of Sub-Mediterranean Downy Oak Landscapes of Southeastern Crimea. *Forests* **2022**, *13*, 1370. https:// doi.org/10.3390/f13091370

Academic Editor: Tim Martin

Received: 26 July 2022 Accepted: 25 August 2022 Published: 27 August 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

maquis and garrigues, respectively, although they differ in some important respects [14]. In contrast to the predominantly sclerophyllous maquis, deciduous species predominate in shiblyak biomes [20]. These forests formed at the furthest extent of their range due to the effect of the subtropical climate, which occurs here due to the circulatory position of the Southern coast of Crimea. Cold air masses from the north tend not to penetrate here, while those from the southern and southwestern directions linger on the south-facing slopes of the Main Ridge. At the same time, Southeastern Crimea is located in the circulatory and insolation shadow formed by the Main Ridge of the Crimean Mountains, which determines the region's high aridity and decreased radiation balance values. Thus, the sub-Mediterranean (or semi-subtropical) climate formed on the territory of the Southeastern Crimea is characterized by warm winters, maximum precipitation during the cold period of the year (late autumn-winter), and hot, mostly dry summers.

In this regard, the study of the water balance of downy oak forests located at the furthest extent of their range represents an important problem, whose solution will further scientific understanding by uncovering individual patterns of the internal organization of such systems. Despite the relatively small areas that they cover, the importance of studying the water balance of downy oak forests in Southeastern Crimea is also due to the ecosystem functions that these forests perform (preservation of moisture reserves in the soil, prevention of negative physical and geographical processes on the slopes, etc.).

The aim of this work is to study the temporal variability and identify trends in the components of the water balance of the sub-Mediterranean downy oak landscapes of the Southeastern Crimea on the example of the Karadag landscape and ecological station. The strong suit of this study is that almost all data are obtained from long-term stationary studies. To achieve this goal, the following tasks were solved: we described in detail the conditions of the forest formation with regard to the influence of landscape components on the water balance of the territory; we revealed the dynamics of water balance components of sub-Mediterranean downy oak landscapes.

Identification of dynamics features and formation of water balance components on the basis of long-term stationary observations, and not only by means of modeling, makes it possible to more accurately identify dependencies and patterns in climate change for Mediterranean territories vulnerable to external changes in environmental conditions in the future.

#### **2. Study Area**

The area of study encompasses the southeastern part of the Crimean Peninsula. The concept of "South-Eastern Crimea" is often encountered in scientific publications. However, no clear boundaries of this region have yet been defined. The analysis of publications shows that the authors most often refer to the South-Eastern Crimea as the territory comprising the municipalities of Sudak and Feodosia. However, this designation is associated not with administrative units, but rather with landscape features. The natural boundaries of the studied area take the form of watersheds. The western border of the district runs along the Goller Ridge watershed between the Raven and Shelen Rivers, including the peaks of Piyakya, Biyuk-Krizh, and Livaz-Kaya. The northern border runs along the main watershed that delimits the basins of rivers flowing into the Black Sea (i.e., in a southerly direction) and the Sea of Azov (in a northerly direction). The northern border runs along the cliffs of Yuvan-Kaya, the Khambal ridge, the peaks of Kukushlu-Oba, Kazas-Olan and Apaly, Tuar-Alan Ridge, and Jady-Kaya Mountain. The northeastern and eastern borders run along the line from Kara-Oba along the Uzun-Syrt and Tepe-Oba ridges to end in the area of Cape Ilya.

The relief of Southeastern Crimea, which is represented by mountain ranges in the central and western parts, is gradually transformed into a plain towards the east. The elevation of the area varies from 0 to 917 m (Figure 1). Forest communities are mainly represented in the most elevated western and central parts of Southeastern Crimea.

**Figure 1.** Territory of Southeastern Crimea.

The distribution of plant communities throughout the region is influenced by its geomorphological structure. Steppe communities within Southeastern Crimea can be described as occupying the lowest areas (average elevation—about 100 m); these are replaced by oak shiblyaks with an average elevation of 230 m, then downy oak forests (average elevation—285 m), rock-oak forests (average elevation—340 m) and hornbeambeech forests (average elevation—470 m). Within the area under consideration, territories with minimal terrain dissection are covered by steppe, psammophytic, and halophytic communities, while territories with the greatest relief dissection are covered by forest communities.

The slopes of Southeastern Crimea are subject to various erosion processes, including ravine formation, which determines the redistribution of forests along the slopes. Snow catchments play an important role in the spatial differentiation of forests. The linear erosion Stream Power Index (SPI) calculated for the Southeastern Crimea shows that most of the area is susceptible to strong erosion processes.

The territory of the Southeastern Crimea is characterized by Mediterranean climatic features. The average annual air temperature within the Southeastern Crimea area, which varies from the northwest to the southeast, ranges from +9◦ to +13◦. Southeastern Crimea experiences between 100 and 300 mm of precipitation during the winter period. The precipitation field decreases from west to east (Figure 2a), with the greatest amount falling in the northwestern part of the study area. In summer, 80 to 160 mm of precipitation falls over the territory of the Southeastern Crimea, with the precipitation field decreasing from the northwest to the southeast (Figure 2b). Both in winter and in summer, the coastal areas of Southeastern Crimea are the most arid. The annual distribution of precipitation varies from 700 to 350 mm in the west-to-east and northwest-to-northeast directions.

**Figure 2.** Average precipitation on the territory of Southeastern Crimea for the winter (**a**) and summer (**b**) periods.

The formation of vegetation cover is affected by the distribution fields of climatic factors. Both the greatest amount of precipitation and lowest temperatures prevail in the most elevated areas of Southeastern Crimea; it is here that hornbeam and beech forests are mainly developed. Along with a decrease in the amount of precipitation, woody vegetation is gradually replaced by shrub and steppe communities. Among other formations, oak shiblyaks and tipchak steppes are characteristic of the areas where the least amount of precipitation will fall.

The density of the river network within Southeastern Crimea decreases from the southwest to the northeast from 0.3 to 0.1 km/km2 [21]. This indicator is indirectly associated with increased climate aridity and decreased precipitation. The watercourses of Southeastern Crimea fully belong to the rivers of the southern macro-slope of the Crimean Mountains. Nevertheless, the river network of Southeastern Crimea is quite complex; in general, the area is characterized by a combination of river valleys, interfluves, and watersheds (Figure 1).

Conditions for the existence of separate small catchments of small gullies and other erosive forms on the coast of Southeastern Crimea are due to the complex rugged terrain, which creates unique topological conditions for the existence of plant communities. For example, local factors disrupt the manifestation of quasi-zonal processes within the valleys of the watercourses of Southeastern Crimea. In addition to oak communities, alder, poplar, and willow are also found throughout the area.

Southeastern Crimea is characterized by a predominance of brown soils (up to 80% of the total area) [22]. Mainly weak and underdeveloped soils are formed on the destruction products of limestones, sandstones, and conglomerates composing individual mountain massifs and ridges, as well as on volcanic rocks of the Karadag massif. In the lower parts of the slopes and at their feet, full-profile and drift soil types are formed. Alluvial (17%) and meadow (2.5%) soil types are commonly encountered in the valleys. Brown mountain–forest soils within the study area have an extremely scanty distribution (about 0.3%).

In certain areas of Southeastern Crimea, vegetation cover has undergone a fundamental transformation. Here, the vegetation is represented either by agricultural crops, or goes

through stages of degradation and restoration, which in turn directly affects the water balance of the area due to the capacity of vegetation to retain and redistribute moisture. The vegetation and soil cover, along with the water balance of soils, are additionally affected by the widespread rooting activity of wild boar, which leads to desiccation of the soil and increased evaporation from its surface [23].

#### **3. Materials and Methods**

In the classical form, the water balance equation [24] can be represented in the following form (Figure 3):

$$\mathbf{x} + \mathbf{y}\_1 + \mathbf{w}\_1 + \mathbf{z}\_1 = \mathbf{y}\_2 + \mathbf{w}\_2 + \mathbf{z}\_2 \pm \Delta \mathbf{u}\_2 \tag{1}$$

where x is precipitation on the surface of the object; y1 is surface inflow of water from outside; w1 is underground inflow of water from outside; z1 is condensation of water vapor; y2 is surface outflow of water outside the object; w2 is underground outflow of water outside; z2 is evaporation; Δu is change in the volume of water within the object (contour).

**Figure 3.** Scheme of the water balance equation. Black arrows show the flows direction.

At the local level of the study, the structure of the water balance becomes an even more complex phenomenon. The structure of the water balance of forests comprises several elements: the flow of water with precipitation and surface runoff from neighboring areas, canopy interception, stem flow, transpiration and evaporation from tree crowns, water flow to the forest litter surface, evaporation from the forest litter surface, infiltration, evaporation of water from the soil, its retention by precipitation (formation of soil moisture capacity and precipitation), surface runoff, subsurface runoff, and water transfer to groundwater runoff.

The study of the water balance components of downy oak forests was conducted under the auspices of the Karadag Landscape and Ecological Station (KLES) of the Karadag Nature Reserve. The KLES is located in the redivision of the major catchment gully (0.6 km long, 17 ha area) on the eastern slope of the Besh-Tash ridge, 1.5 km from the village of Kurortnoe. There, the precipitation and soil moisture parameters are measured in a meteorological site located in an open area in the lower part of the catchment. A precipitation collector and soil lysimeters are also installed in the middle part of the catchment area under the canopy

of the forest. Observations of soil moisture are carried out at 30 points located in areas with forest, shrub, and steppe vegetation. Monitoring of the runoff is carried out on two runoff sites located within the KLES area. Figure 4 shows the network of monitoring stations.

**Figure 4.** Area covered by the Karadag Landscape and Ecological Station (KLES) and monitoring points.

The KLES area is representative of the entire Karadag Nature Reserve, since almost all types of plant communities are represented here. The vegetation cover is characterized by a strong mosaicism, which is due, among other things, to the relief dissection. More than 20% of the KLES area is occupied by forest communities. They are mainly represented by a downy oak formation (*Querceta pubescentis*) with the addition of *Acereta campestrii*, *Pistacieta muticae*, and *Pineta pallasianae*. Shrub communities, which occupy about a tenth of the study center area, are represented by formations of thorny shrubs—*Paliureta spinae-christi*, *Rosaeteta corymbiferae*, *Pruneta spinosae*, and *Rubeta taurici*. Herbaceous communities are quite diverse and occupy more than 60% of the hospital territory [9].

In order to obtain values for building evapotranspiration maps, data from the open database of the ERA5 reanalysis of the Copernicus Climate Change Service project [25], implemented by the European Center for Medium-Range Weather Forecasts (ECMWF), were used. As a result, the average monthly data for 1978–2021 were obtained. The data are presented in the netCDF4 format and cover the territory in question with a regular grid with a cell size of 0.25◦ × 0.25◦.

#### **4. Results and Discussion**

The water balance of any territory depends on climatic factors and the nature of the underlying surface. The literature provides various values of the average annual precipitation falling on the territory of the Karadag reserve. On average, 357 mm of precipitation falls on Karadag (for the period from 1930 to 1980) [26]; according to other data, 389 mm of precipitation (for the period from 1920 to 2006) [27], 501 mm of precipitation (for the period from 2000 to 2011), and 467 mm of precipitation (for the period from 2012 to 2018) according to [28].

According to our observations, an average of 448 mm of precipitation fell annually on the territory of Karadag from 2010 to 2020. This is significantly less than in a typical Mediterranean region. Due to the presence of downy oak forests in Northeastern Crimea, environmental conditions can be considered as transitional from optimal to pessimal. For example, within the limits of downy oak forests growing about 15 km north of Montpellier in southern France at an altitude of 250 m, the average annual precipitation is 1311 mm [29]. In Italy, downy oak forests receive about 850 mm of precipitation; in the area of the city of Evora in Portugal, the corresponding figure is 665 mm [30].

In terms of the interannual quantitative dynamics of incoming precipitation, a wide range of values is observed: in 2020, an average of 275 mm of precipitation fell on Karadag, while in 2010, the corresponding amount was 715 mm. Thus, the amplitude reaches 440 mm of precipitation. In 2010–2020, there is a tendency towards a reduction in the amount of precipitation: 2019 and 2020 are both characterized by extremely low average annual precipitation values.

In the intra-annual dynamics, a constant winter maximum of precipitation is observed within the downy oak forests. However, in some years (2013, 2014), absolute precipitation maxima were observed in the study area in June (45% of the annual precipitation), which were catastrophic for the environment. July and August are characterized by stable arid conditions, with an average of 31 mm precipitation. The winter maximum is characterized by a predominance or uniform distribution of precipitation during the late autumn and winter months. The average amount of precipitation in December is 45 mm; in January, 48 mm; in February, 33 mm. Spring and autumn precipitation minima are characterized by a decrease in the amount of precipitation: on average, 27 mm of precipitation falls in spring within the downy oak forests, while in autumn, average precipitation is 34 mm.

In downy oak communities, retention of precipitation in crowns is around 24% of precipitation, comprising an average of 104 mm. Considering their relatively small amount of precipitation, the forest communities of Karadag can retain up to 100% of precipitation [28]. At the same time, there is an intra-annual dynamic to this process, associated both with the vegetation stages of plants and with the intensity of precipitation. Thus, the maximum values of precipitation retention by crowns are typical for the second half of spring to early summer due to plants achieving a maximum of green phytomass during this period. For example, in July, an average of 9 mm (about 41%) of precipitation of precipitation is retained. The retention of precipitation by tree crowns is also affected by the precipitation intensity. In [28] it is indicated that up to 20% of precipitation is retained when more than 20 mm precipitation falls on Karadag, while when precipitation is less than 2 mm, over 45% is retained. Data analysis shows that the increase in precipitation observed in Southeastern Crimea is formed due to summer heavy rainfalls [16]; this in turn leads to most of the water penetrating under the canopy of the forest, with little retention in the crown. In addition, large drops characteristic of heavy rainfall are poorly retained on the leaves, quickly draining to the surface of the forest floor.

A significant role in the water balance structure of downy oak forests is played by stem flow. This effect is interesting due to its concentration of water flow directly under the roots of trees. However, under conditions of sloping terrain, this process can intensify sheet-, drip-, and fine-jet erosion of soils, exposing the roots of trees and accelerating the processes of badland formation. In [9], a methodology for assessing stem runoff in the territory of Karadag is given; however, quantitative characteristics of the experimental results are lacking. A.A. Klyukin in [31] notes that in an oak forest with a canopy density of 0.8, growing on a slope with a steepness of 20–25◦ and a length of 100 m, the average accumulation of erosion near tree trunk is about 0.004 mm/year. At the same time, A.A. Klyukin notes that the average rate of erosion can reach 2.0–3.9 mm/year on steep and

short slopes that are adjacent to river valley beds and talvegs of temporary watercourses, accompanied by exposure of the parent rock. Thus, such processes are characteristic of steep slopes. On slopes of medium and low steepness, such processes are not significantly manifested, with the retentiveness of the soil being provided by an herbaceous layer. A.A. Klyukin in [31] notes that the trunks of an adult oak can run off 5–6 L of precipitation per year; in our opinion, this figure is significantly overestimated for the forests under study. The more objective data in [7] indicate that runoff from oak trunks reaches 3%–5% of the annual precipitation layer, which is equivalent to approximately 20–40 mm/year. Stem runoff is also actively affected by the roughness of the bark of tree trunks, which demonstrates that the real area of the runoff is larger than the ideal cylindrical shape of the trunk. In oak forests, this indicator, according to [9], reaches 1.2, allowing them to be attributed to medium rough type. By comparison, stem runoff in oak forests in Spain accounts for slightly more than 10% of precipitation [32].

According to measured data on sample trees, throughfall accounts for about 75% of the total precipitation amount. Comparable results were obtained by V.A. Bokov's expeditions in this area in 1997 and 1998. [9]. According to his research, a dense downy oak forest intercepts about 25%–35% of precipitation during the growing season from May to October, which is about 103 mm in a wet summer (1997) and 55 mm in a dry one (1998). However, this moisture, which enters the soil surface more gradually than heavy rainfall, is generally well-retained by the forest floor, helping to form its moisture reserve. This corresponds to around 28 and 35%, respectively, of the precipitation depth that fell during the growing season, or 18 and 16% of the annual precipitation layer. During winter, when vegetation is absent in deciduous forests and there is no foliage on the trees, precipitation almost completely penetrates under the canopy of the forest, only being slightly delayed by tree branches. According to the results of measurements carried out in the winter season, precipitation retention of slightly more than 20% does not exceed the annual average. However, given that the winter months account for the maximum precipitation and taking into account the seepage through the crowns, this indicator is much higher than the summer maximum and accounts for no more than 15 mm of precipitation. These data are partially confirmed in the work [14], where it is indicated that during the period from 2000 to 2008, 162 mm was retained within the downy oak forests of Southeastern Crimea, which is approximately 30% of the amount of precipitation. At the same time, observations show that heavy rains are not so characteristic of the winter precipitation maximum, resulting in the possibility of a satisfactory soaking of the forest litter and soil. M.A. Kochkin notes [7] that the forest floor of oak forests, consisting of leaves and dead herbaceous vegetation, absorbs and retains water from precipitation, thereby reducing surface runoff.

For the period from 2010 to 2020, there has been a significant decrease in soil moisture: in average annual terms, by about 1.6 times. The inter-seasonal dynamics of decreasing soil moisture shows that almost all seasons of the year are characterized by a drop in average relative humidity values close to the average annual value, with the exception of autumn, when relative humidity decreases by 1.4 times. Considering the interannual dynamics, it is important to highlight a significant reduction in soil moisture in December (by more than 2.5 times), as well as minor reductions in September. The maximum relative values of soil moisture are experienced during the winter months, as well as in March, when the average value does not fall below 25%, while the minimum values occur at the end of summer and beginning of autumn (August and September), when the relative soil moisture values drop to 15%–18%.

During the period from 2010 to 2020, evapotranspiration within the downy oak forest community averages 450 mm. These values are generally equivalent to the evapotranspiration of oak forests in the Mediterranean region: 343 ± 37 mm in the Mediterranean region as a whole [30], 458 mm in the province of Tarragona in Spain [32], 823 mm in the eastern Pyrenees [33]. Maximum evapotranspiration values in the Mediterranean can reach 1300 mm per year [34].

Analyzing the interannual evapotranspiration dynamic within the oak forests of Karadag, its values can be seen to form a steadily increasing trend (Figure 5).

**Figure 5.** Average annual values showing the relationship between evapotranspiration and precipitation in the downy oak forest in the period 2010–2020.

When considering the seasonal evapotranspiration values, the fact of the predominance of precipitation over evapotranspiration in the winter and autumn periods becomes clearly evident. During winter, the average evapotranspiration value is 11 mm, which is approximately 26% of the amount of precipitation. In autumn, the average evapotranspiration is 22 mm, which is approximately 63% of the average amount of precipitation for this time of year. In the remaining seasons of the year, there is an excess of evapotranspiration values over precipitation. In spring, the average evapotranspiration of 56 mm is 2 times higher than the average amount of precipitation for this season; in summer, the equivalent amount is 67 mm, which is 1.5 times more than the average amount of precipitation at this time of year. At the same time, a steady tendency towards increased average annual values of evapotranspiration is evident at all times of year.

When considering the distribution of evapotranspiration by month, the highest values of evapotranspiration are characteristic of May, June, and July, when the average values of evapotranspiration exceed 70 mm. The lowest values are experienced for December, January, and February, when average evapotranspiration does not exceed 15 mm. In spring, there is a gradual increase in evapotranspiration values, while in autumn there is a gradual decrease. Nevertheless, data analysis also demonstrates a significant increase in evapotranspiration occurring in recent years during late summer and early autumn.

A significant role in the spatial differentiation of downy oak forests is played by accumulations of snow. Around 5% of precipitation falls in the form of snow in the Karadag area [26]. According to [9], in the Southeastern Crimea in the area of Karadag and Echkidag, between 50% to 90% of snow can be blown off the windward northeastern slopes. Although winter precipitation is generally not retained by tree crowns due to the absence of foliage, it can be carried as a result wind transport. For example, in 2019, a layer of snow with a 15 cm high was formed as a result of wind activity. For the period from 2011 to 2020, the average annual precipitation in solid form was 38 mm, accounting for approximately 9% of the total. However, in terms of the interannual dynamics of the indicators under consideration, the years with the minimum amount of solid precipitation (2013, 2020) can be distinguished, when the amount of precipitation in solid form did not exceed 10 mm, as well as the years of maximum solid precipitation—for example, 2015 and 2016—when the corresponding amount was more than 60 mm. Despite the duration of snow in Southeastern Crimea lasting no more than a week, the phenomenon

of snow catchments is very characteristic; where the accumulated snow is blown onto the slopes of the southern exposures, more hydrophilic conditions are formed on these slopes than those with northern exposures. For example, within one of the gullies on the eastern slope of the Besh-Tash Range, the slope of the southern exposure is characterized by better moisturization than the opposite slope. This is due to the extensive snow collection adjacent to the slope of the south-southwestern exposure, from which the snow is blown by northeasterly winds to the slope of the south-southwestern exposure (Figure 6) [35]. In this regard, the slope of the south-southwestern exposure is better forested and characterized by greater erosion indentation as compared with the opposite slope of the north-northeastern exposure and slopes of the south-southwestern exposure, where areas with heavily washed soil cover are noted.

**Figure 6.** Snow-covered slopes of the south-southwest exposure of the gully on the eastern slope of the Besh-Tash ridge [35].

The formation of surface runoff depends on a whole complex of factors. A slope runoff characterized by insignificant indicators is identified on the basis of observations carried out at the runoff sites. In most cases, runoff is caused either by melting snow or heavy precipitation. In 1996–1997, the average runoff was approximately 0.16 L/km<sup>2</sup> [9]. The average value of the runoff from 2010 to 2020 recorded by the measurement results was 0.04 L/km2. The highest average runoff values, which were recorded in 2010, amount to 0.09 L/km2, while the lowest figures recorded in 2011 and 2020 were slightly less than 0.02 L/km2. Thus, these values have no significant effect on the final equation of the water balance of downy oak forests. A decrease in the values of surface runoff in the interannual dynamics is noted for the period under review.

Horizontal precipitation (condensation from fog, frost, frost, ice, etc.) plays a prominent role in the formation of the water balance of downy oak forests. Although we unfortunately did not have the opportunity to conduct real measurements of their contribution to the formation of the water balance of forests, according to the data [36], their contribution can reach 20%–25% of the annual precipitation layer. From September to March, no more than 20 days with foggy conditions can be observed within the downy oak forests; on an annual basis, the number of foggy days varies from 20 to 40 [37]. According to the Feodosiya weather station [8], there is an average of 24 foggy days per year. In 2020, 36 days with fog were recorded within the downy oak forest, of which 75% occurred during the winter period, significantly affecting the incoming part of the water balance.

All the processes described above take place against a background of the displacement of steppe communities by forest communities on the upper parts of the slopes (Figure 7). The growth of heavy rainfall in summer leads to a rapid flushing of water from the upper parts of the slopes, which are characterized by petrophytous steppe communities, along with an increase in soil moisture in the lower parts of the slopes, which creates conditions for forest communities to rise up the slope, reducing the number of slope microzones. This phenomenon is clearly visible when calculating the normalized difference vegetation index

(NDVI) (Figures 7 and 8). When analyzing the average NDVI values for August, which were obtained from Sentinel-2 satellite images in the period from 2016 to 2018 and from 2019 to 2021, an increase in the values can be clearly observed, along with a corresponding increase in the quantity of green phytomass. Figure 8 shows that forest vegetation increases up to 158 m on the profile AB, with the NDVI values not significantly changing from one period to the next. Conversely, above the 158 m mark, where there is initially less woody vegetation, and in August, respectively, less green phytomass, there is an increase in NDVI values and, accordingly, the amount of green phytomass.

**Figure 7.** Comparison of NDVI values in 2016–2018 (**a**) and 2019–2021 (**b**). AB—slope profile.

**Figure 8.** Change of NDVI values for the period from 2016 to 2018 and from 2019 to 2021 from the foot of the slope to the top along profile AB.

#### **5. Conclusions**

When studying the forests water balance, it is extremely important to describe in detail the conditions its formation: geological structure, relief, hydrological features of the territory. It is necessary to give a description, characterizing their influence on the spatial differentiation of forests. So, summarizing the above, the slopes of the Southeastern Crimea are subject to ravine formation, which determines the redistribution of forests along the slopes. A significant role in the spatial differentiation of downy oak forests is played by accumulations of snow. In general, the Southeastern Crimea is the territory of river valleys and interfluves; it is a mudflow area. The passage of mudflows in the strongest way transforms the appearance of forests. The mechanical composition and gritty consistence of soils determine the water regime of forest communities. In the study area, we are dealing with very lithosolic soils formed on the weathering products of flysch, limestone, marl, intrusive, and effusive rocks. It is extremely diverse in terms of parent rock material, and, accordingly, in terms of forest conditions. Forests of the Southeastern Crimea have a different appearance and combines with different types of steppe communities. The study area includes forests, woodlands, and shrub communities. A vivid example of the influence of the animal world on the formation of the water balance of forest ecosystems is the burrowing activity of the wild boar, which is ubiquitous and leads to the drying up of the soil, increasing evaporation from its surface.

Based on field research conducted in the Karadag Nature Reserve and analysis of literature data, the main components of the water balance of the downy oak forests of Southeastern Crimea are considered for the period from 2010 to 2020.

The average annual precipitation on the territory of Karadag for the analyzed period was 448 mm; in recent years, a decrease in the amount of precipitation has been recorded. On average, the crowns of downy oak communities retain up to 24% of precipitation. In terms of seasonal dynamics, the maximum retention of precipitation (up to 41%) occurs in the period from late spring to early summer; this is due to the achievement of maximum volumes of phytomass during this period.

According to the literature data, stem runoff intercepts 3%–5% of the annual precipitation; on steeper slopes only, this can lead to negative consequences in the form of planar soil flushing. Runoff formed from the crowns of trees reaches the surface of the soil more gradually to forms the moisture reserve of the forest floor. During the study period, soil moisture decreased significantly by 1.6 times. The maximum relative soil moisture values are experienced during the winter months and in March, when the average value does not fall below 25%. The features of the relief of the Southeastern Crimea created conditions for the redistribution of solid precipitation (snow) from the windward slopes to form snow catchments. Due to this effect, the slopes of the southern exposures are characterized by better moisturization than the opposite slopes of the northern exposures.

The evapotranspiration values within the downy oak forests approximately coincide with the values of this indicator in the Mediterranean region to average 450 mm per year. The interannual dynamics demonstrates a positive trend towards increased evapotranspiration values. A significant increase in evapotranspiration observes in late summer and early autumn.

Observations of slope runoff conducted at runoff sites suggest that the final value of the water balance is not significantly affected by this factor. The average value of the slope runoff was 0.04 L/km2; in terms of interannual dynamics, a decrease in this value is recorded for the period under review.

All the analyzed features of the water balance elements form conditions conducive to the displacement of steppe communities by forest species, which is confirmed by the results of field and remote studies.

**Author Contributions:** Conceptualization, R.G.; methodology, R.G.; validation, R.G.; formal analysis, V.T., T.G. and M.S.; investigation, V.T., T.G.; writing—original draft preparation, V.T., T.G., M.S. and R.G.; writing—review and editing, R.G.; visualization, T.G. and V.T.; supervision, R.G.; project administration, R.G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Russian Science Foundation, grant number 22-27-00579.

**Data Availability Statement:** Not accessible.

**Acknowledgments:** The authors are grateful to Kelip A.A. for technical support of the study.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland www.mdpi.com

*Forests* Editorial Office E-mail: forests@mdpi.com www.mdpi.com/journal/forests

Disclaimer/Publisher's Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Academic Open Access Publishing

mdpi.com ISBN 978-3-0365-9637-2