Field-Measured Hydraulic Traits and Remotely Sensed NDVI of Four Subtropical Tree Species Showed Transient Declines during the Drought–Heatwave Event

Abstract

Unpredictable drought–heatwave events occur frequently worldwide, causing low water availability (drought) and high temperatures (hot), with consequences for forest decline and mortality. Our knowledge of the potential instantaneous reactions and subsequent recovery of water-related physiological processes and vegetation indices in hot drought events remains unclear. Here, we investigated how the 2022 summer drought–heatwave event in the subtropical regions of China affected hydraulic traits and NDVI values in the forests of four common subtropical tree species. During the hot drought, the NDVI values of all four forests decreased (−31%~−23%), accompanied by leaf scorch and tree crown dieback. Among the four species, a hot drought event caused an instantaneous descent in hydraulic conductivity (K_s, −72%~−31%), stomatal conductance (g_s, −94%~−50%), and midday water potential (−40%~−169%), with severe drought-induced stem xylem embolism. A trade-off was found between resistance and resilience in hot-drought-induced hydraulic dysfunction, as species with lower declines in K_s and g_s during the hot drought had a shorter recovery in the post-stress phase. This study highlights that the 2022 hot drought event had severe negative instantaneous impacts on the forests of four subtropical tree species, which were reflected both in water-related physiological processes in the field and in remote sensing data from satellites.

1. Introduction

During the last 50 years, extreme weather events have become more frequent and intense due to climate change [1,2,3]. According to the Sixth Assessment Report of the IPCC, the global average temperature in 2011–2020 increased by about 1.09 °C compared with 1850–1900 [4]. In the future, the frequency and intensity of extreme weather events will increase [5]. Drought–heatwave events—combinations of extreme weather events imposed by drought and high-temperature heatwaves [6]—are particularly striking [7,8,9,10,11]. During the past 20 years, frequent drought–heatwave events have led to an increase in tree mortality worldwide [12,13,14] and catastrophic impacts on forest structure, function, health, ecosystem services, and sustainability [15,16,17,18,19,20]. Moreover, drought–heatwave events will seriously restrict regional forest ecology development and damage the social and ecological environment [21]. In the summer of 2022, an extreme drought–heatwave event swept the entire Northern Hemisphere, with many countries experiencing hot drought weather, including China [22]. This extreme compound dry–hot climate event occurred unpredictably and intensely and covered a wide area in subtropical China; it even broke the record for the highest temperature during the same period in several subtropical regions [23,24,25,26,27]. Under the combined action of extreme compound dry–hot climate events and urban heat islands, subtropical forests in Eastern China have suffered double damage, and withering has occurred in a large area. Tree species growing in this area have suffered from destruction imposed by high temperatures and drought during the growing season. Therefore, a serious injury has occurred, which needs to be recorded.

Much less attention has been given to instantaneous reactions under hot drought events and the subsequent recovery of tree species processes in the field, despite the fact that many trees will be increasingly exposed to heatwaves with global-change-type droughts in the near future [28,29,30,31]. During extreme hot drought events, high temperatures and drought will occur at the same time [27], resulting in increased difficulty for trees to obtain water [14] and a negative water potential, leading to the production of cavitation in xylem water transport, disruption of the xylem hydraulic balance, discoloration or even premature abscission of leaves [32,33], and dehydration of stems [34]. At the same time, under the stress of high temperatures and drought, the chlorophyll in the leaves will be destroyed, photosynthesis will be disordered, and death may even occur when the hydraulic balance and photosynthesis are irreversibly damaged [35,36]. Moreover, the occurrence of such extreme weather will also reduce the NDVI of a forest by destroying its growth state [37]. In the severe context of such climate change, the impact of drought and heatwaves on trees will be more prominent. Temperature and environmental factors have an important impact on the growth and survival of plants, and high temperatures usually increase the water vapor pressure deficit (VPD), which in turn increases the transpiration loss of leaves and plants [38]. In the case of a high VPD, the leaves of trees reduce their temperature by increasing the intensity of transpiration to avoid burns [39]. With the increase in the duration and intensity of high temperatures, the VPD in the atmosphere will be continually aggravated, which will destroy the tree water balance, seriously affect tree life activities, and even lead to the death of trees [40]. Leaf photosynthesis also changes with atmospheric temperature [41]. Under high-temperature conditions, drought caused by long-term soil moisture deficiency can cause plant xylem embolism and further inhibit growth and photosynthesis [42].

Indeed, tree responses to high temperatures and drought events are a hot topic that scholars are paying attention to [43]. Some studies have looked at the physiological response of one aspect of a forest to extreme dry–hot climate events [44,45,46]. Other studies have explored the overall impact of tree response results on the ecological community and ecosystem of an entire region based on the long-term results of forest responses to extreme dry heat climate events [10]. However, the real-time impact of extreme combined heatwave–drought climatic events on trees and the associated instantaneous physiological response mechanisms are not fully understood—information that is crucial for guiding reasonable protective measures during hot drought stress. Otherwise, studies that focus on the instantaneous effects of drought and heatwave events on hydraulic and photosynthetic processes mostly depend on controlled laboratory simulation experiments conducted on seedlings under reduced water and increased temperature conditions [47,48]. These studies show that, on a small scale, temperature and drought factors have a huge destructive effect on the hydraulic balance of tree seedlings, and this effect lasts for a long time, meaning the seedlings cannot recover quickly in a short period of time and eventually die. Due to the unpredictability of drought–heatwave events and the differences in complex environmental conditions between the field and laboratory, as well as the discrepancy in the resistance and recovery ability of trees and seedlings to hot drought stress [41,49,50], the investigation of the instantaneous response of trees during extreme drought and high-temperature events and the degree of recovery in subsequent post-hot drought events for tree species in the field requires further research and exploration.

In addition, in view of the important value of remote sensing satellite imagery as a potential “gold mine” [51,52] and the importance of the normalized vegetation index (NDVI) in reflecting vegetation growth vitality [7], monitoring climate change, and environmental research [53], many studies have obtained different forest growth spectral indicators through remote sensing monitoring to explore the growth state changes in forest systems before and after high-temperature and drought events. Remote sensing image analysis undoubtedly makes up for the shortcomings of forest ecosystem research on larger spatiotemporal scales. However, due to typical regional differences in forest ecosystems, the distribution of major forests varies markedly in different parts of the world. Therefore, even though remote sensing analysis is an important research method, the small-scale determination of forest physiological indicators in a particular area and large-scale forest spectral analysis have not been effectively combined; in many studies, the measurement of tree physiological indicators was conducted alone to obtain the growth state of trees in one region [28,54], and other studies have applied remote sensing analysis to monitor another region [55]. Although many studies have begun to combine these two scales’ characteristics, limited physiological indicators have been recorded, and a mismatch still exists between large-scale vegetation indices and physiological process traits, meaning that the combination of these two scales’ characteristics requires more accurate recording and analysis [56,57].

This study was conducted against the background of frequent and unpredictable global high-temperature and drought events, the actual background of the great impact of forests in various places, the relatively individual research methods in the field of forest ecology, the failure to effectively combine large-scale research methods and small-scale measurement methods, and the relatively weak research on the difference in the instantaneous response degree of different tree species to hot drought stress. In order to guide forest management under hot drought stress in the near future [58], we combined remote sensing monitoring with field measurements of various physiological traits of forests with different tree species to explore the resistance and resilience of different trees to compound dry–hot events and the interconnection between various indicators through the monitoring of the NDVI values of vegetation on a large scale and the measurement of hydraulic and photosynthetic traits of each tree species on a small scale.

In our study, measurements of remote sensing images were recorded at the same time and site in three stages, focusing on the physiological traits of tree species in the field, such as hydraulic and photosynthetic traits. Here, we recorded stem hydraulic conductivity, midday water potential, leaf water content, etc., as well as leaf gas exchange traits and chlorophyll content in three time periods, i.e., before (June 2022), during (July 2022), and after (early September 2022) the extreme dry–hot climate event. In addition, by comparing the changes in the growth health status of different tree species before and after the drought–heatwave event, this study evaluated the resistance of different tree species to compound dry–hot events and their subsequent resilience, which may provide guidance for forestry construction under climate change. Moreover, we also analyzed the changes in various environmental meteorological factors, the forest NDVI, and physiological characteristics with the aim of providing a mechanistic explanation for the different coping strategies among the species during the drought–heatwave event.

Specifically, we aimed to answer or validate the following questions and hypotheses:

(1) During the extreme drought–heatwave event, did the NDVI of the four forests in this study decrease? Was the decline in the NDVI instantaneous?;

(2) Furthermore, in the subsequent post-drought–heatwave event phase, did the NVDI of the four forests recover instantaneously? Additionally, what was the percentage of recovery of the NDVI value of these four forests compared to before the event?;

(3) During the extreme dry–hot event, the physiological activities of the four tree species were reduced instantaneously. Compared with the gymnosperm species MG and TD, the hydraulic architecture of the angiosperm species KP and CC showed higher resistance to the dry–hot event with a smaller reduction in hydraulic conductivity and leaf water potential.

2. Materials and Methods

2.1. Study Site and Plant Species

This study was conducted at the Yangtze River Delta National Observatory of Wetland Ecosystem, Shanghai (121°12′ E, 31°15′ N), located at the junction of Jiangsu, Zhejiang Province, and Shanghai, Eastern China. The area has a subtropical marine monsoon climate, and the mean annual rainfall is 800–1500 mm, mostly occurring in spring and summer. The research field in this study included four sites: Dalianhu, Liantang, Nanyuewei, and Zhangma Village, which are typical areas for the distribution of major subtropical afforestation tree species in the Yangtze River Delta region (Figure 1). The four tree species included in this study were Koelreuteria paniculata Laxm., Cinnamomum camphora (L.) Presl, Metasequoia glyptostroboides Hu and W. C. Cheng, and Taxodium distichum var. imbricatum (Nuttall) Croom. As dominant afforestation tree species widely distributed in the eastern subtropical region of China, these four tree species are also commonly distributed in our research field. Among them, the dominant tree species distributed in the Dalianhu research field are Koelreuteria paniculata Laxm. and Cinnamomum camphora (L.) Presl; the dominant tree species distributed in Nanyuewei is mainly Cinnamomum camphora (L.) Presl; and Metasequoia glyptostroboides Hu and W. C. Cheng and Taxodium distichum var. imbricatum (Nuttall) Croom are widely distributed in the Liantang and Zhangma Village research sites. These four research sites suffered from severe drought and heatwave events in 2022, and the four sites had measured temperatures above 40 degrees and air humidity below 45% in the field for several days, with typical high temperatures and drought conditions. These sites are prominent representatives for the study of subtropical trees in Eastern China.

2.2. Meteorological Data Collection and Tidying

Based on the hourly and daily real-time meteorological observation data of the China Meteorological Agency (CMA), this study collated and calculated the average monthly temperature, precipitation, relative air humidity (RH), soil temperature (T_s), volumetric water content (VWC), and soil moisture (M_s) in Shanghai for 2022. We calculated the daily saturation vapor pressure difference (VPD) by calculating the daily maximum temperature and daily relative air humidity (RH) using Equation (1), and then obtained the monthly average. The saturation vapor pressure difference (VPD) was calculated as follows [59]:

$$\mathrm{VPD}=\mathrm{a}\times {\mathrm{e}}^{\frac{\mathrm{b}\times {\mathrm{T}}_{\max .\mathrm{d}}}{\mathrm{c}+{\mathrm{T}}_{\max .\mathrm{d}}}\times \left(1-\frac{\mathrm{RH}}{100}\right)}$$

(1)

where VPD (kPa) is the saturation vapor pressure difference, e is a natural constant, T_max.d (°C) is the maximum daily temperature, and RH (%) is the relative air humidity. The constants take the following values: a = 0.6108, b = 17.27, and c = 237.3.

2.3. Remote Sensing Monitoring

We obtained images captured by the Landsat 8-9 Operational Land Imager (OLI) and Thermal Infrared Sensor (TIRS) Collection 2 Level-1 15 to 30 m from the official website of the United States Geological Survey (USGS), available at EarthExplorer (usgs.gov), from May to October 2022. After that, we fused the multispectral images and panchromatic band images from the Landsat 8-9 remote sensing data with the Gram–Schmidt Pan Sharpening tool to fuse them into 15 m resolution remote sensing images and performed radiometric correction and FLAASH atmospheric correction on these remote sensing images to reduce the spectral reflectance error of the images.

Using the Landsat 8-9 reflectivity band, we calculated the normalized vegetation index (NDVI) using Equation (2), and after cropping the remote sensing images of the four forest sample areas, we obtained the histograms of the vegetation index at different times in different areas. The remote sensing image analysis and image processing in this study were performed using ENVI 5.3 (Exelis Visual Information Solutions Inc., Tysons Point, VA, USA) and ArcMap 10.8 (ESRI Inc., Redlands, CA, USA).

$$\mathrm{NDVI}=\left(\mathrm{NIR}-\mathrm{R}\right)/\left(\mathrm{NIR}+\mathrm{R}\right)$$

(2)

where R and NIR are the digital numbers (DNs) in the red and near-infrared channels, respectively.

2.4. Water Potential Measurements

We measured the midday (Ψ_midday) and predawn (Ψ_predawn) water potentials on intermittent sunny days from July to September in 2022 with a pressure chamber (PMS 1505D-EXP, Albany, OR, USA). We took leaflet samples from six individuals of each species under the same conditions for the predawn and midday water potentials before sunrise (04:00–06:00 a.m.) and at midday (12:00–14:00 p.m.), respectively. All the samples were cut from trees after being marked randomly, wrapped with aluminum foil, sealed with plastic bags containing moist paper towels, and kept in a cooler until we conducted the water potential measurements in the laboratory within an hour after excision [60].

2.5. Hydraulic Conductivity

In the early morning, we collected 10 branches that were about 2 m in length from various individuals of each species to carry out the hydraulic conductivity measurements. Once the branches were cut from the trees, they were quickly placed in a bucket of water and then cut under water again with 5 cm removed. The branches were transported to the laboratory at once, with the cut end submerged in water and the other end of the branch covered with wet black plastic bags. We chose an unbranched stem segment that was about 20 cm in length for angiosperms and 7 cm in length for gymnosperms, with a 0.5~1.0 cm diameter, from the large branches collected for the hydraulic conductivity measurements. Wet plastic bags were used to keep the leaflets terminal to each of the stem segments for the hydraulic conductivity measurements in order to measure the leaf areas later.

To prevent vessel clogging from wound reactions, we used a sharp razor blade to shave both ends of the segments each time before these segments were connected to a tubing apparatus filled with filtered, degassed 20 mmol L⁻¹ KCL solution for the hydraulic conductivity measurements. We generated a hydrostatic pressure that caused the KCL solution to flow through the segments under a 50 cm hydraulic head.

Hydraulic conductivity (kg m s⁻¹ MPa⁻¹) was calculated as follows [61,62,63]:

$$K_{\mathrm{h}}=\frac{J_{\mathrm{V}}}{\Delta \mathrm{P}/\Delta \mathrm{L}}$$

(3)

where J_v is the flow rate through the segment (kg s⁻¹) and $\Delta$P/$\Delta$L is the pressure gradient across the segment (MPa m⁻¹). Briefly, 0.1% methyl aniline blue dye was perfused into one end of the segment for the K_h measurement under the same 50 cm pressure head so that the dye-stained cross-sectional area could be smoothly calculated using ImageJ 1.48v software (US National Institutes of Health, Bethesda, MD, USA) to determine the sapwood area (SA). Then, the stained cross sections were photographed from both sides of each segment through an optical scanner (HP Scanjet G3110, Hewlett-Packard Development Co., Beijing, China). The stem-specific hydraulic conductivity (K_s; kg m⁻¹ s⁻¹ MPa⁻¹) was calculated as the ratio of K_h to SA.

2.6. Measurement of Native Embolism

After the initial measurement of K_h, stem segments were shaved at both ends with a sharp razor blade and then flushed with filtered, degassed 20 mmol⁻¹ KCL solution under 0.1 MPa pressure for 20 min to remove air bubbles in the xylem. The maximum hydraulic conductance (K_h-max) was then measured.

The percent loss of hydraulic conductivity (PLC) was calculated as:

$$\mathrm{PLC}=\frac{100\left(K_{\mathrm{h}-\max }-K_{\mathrm{h}}\right)}{K_{\mathrm{h}-\max }}$$

(4)

2.7. Measurement of P₅₀ and Safety Margin

The bench-top dehydration method was used to construct vulnerability curves for the four tree species in this study. Sampled large branches from various individuals for each species were covered in at least two big black plastic bags within wet wipes, which allowed samples to dehydrate slowly across a range of water potentials. Before the measurement, sampled branches were bagged for at least 1 h to equilibrate the water potential across the whole branch. In order to obtain a high water potential, branches were rehydrated with the cut side under water and the other side in two black bags for 2 h. A dozen stems with healthy leaves from the large branches were then removed under distilled water, with approximately 10–30 min between each cut, which was determined by the water potential. Leaves from the same stem were used for water potential measurement. The stems were excised, and the native embolisms shown in PLC were obtained as described above. PLC at a range of water potentials was captured to construct a vulnerability curve. P₅₀ is the water potential at 50% of maximum hydraulic conductivity. The safety margin was then calculated as the difference between Ψ_midday and P₅₀ for each tree species.

2.8. Leaf Gas Exchange

We measured the gas exchange rates of mature sun-exposed leaves from each species with an LI-6400 photosynthetic system (LI-COR Inc., Lincoln, NE, USA). All the measurements were performed between 9:00 and 11:30 am on sunny days from July to September 2022 in Shanghai. We chose three mature sunlit leaves from six individuals and measured them in a leaf cuvette with the temperature maintained at 20–25 °C, a CO₂ concentration of 350–400 µmol mol⁻¹, and an ambient humidity between 65 and 75%. The photosynthetic photon flux density in the leaf chamber was fixed at about 1750 µmol m⁻² s⁻¹. The measurements were averaged across individuals for further analysis [60].

2.9. Leaf SPAD Value and Relative Water Content of Leaves

In this study, we used a SPAD-502 chlorophyll analyzer (SPAD-502 Plus, Minolta, Japan) to determine chlorophyll content. Before the measurement, we used the standard plate provided by the manufacturer for the functional measurement to ensure the normal function of the instrument. During the measurement, the leaves of each tree species were divided into left and right regions according to the leaf veins and then equally divided into the tip, middle, and base regions according to the distance between the parts of the blade and the petiole. In this way, the whole leaf was divided into six parts. Finally, from the right leaf base as the starting point, the SPAD values of each region were determined in a counterclockwise order, and then 6 SPAD values were obtained. The SPAD value of the whole blade was the average of the 6 values.

In order to obtain the leaf relative water content (RWC), we first recorded the fresh weight of the leaves. Then, we placed them in distilled water for 5–6 h until they were saturated, wiped off the surface water, and weighed them to obtain the saturated weight. Finally, they were dried in an oven at 65 °C to a constant mass to determine their dry weight. The RWC of the leaves of the different tree species was calculated by taking the average value of six repetitions [64,65].

RWC was calculated as:

$$\mathrm{RWC}\left(\%\right)=\frac{\left(\mathrm{fresh} \mathrm{weight}-\mathrm{dry} \mathrm{weight}\right)}{\left(\mathrm{saturated} \mathrm{weight}-\mathrm{dry} \mathrm{weight}\right)}\times 100\%$$

(5)

3. Results

3.1. Meteorology Conditions during the Drought–Heatwave Event in Summer 2022

The drought–heatwave event in summer 2022 (Figure 2a,e) was characterized by an average T of 31.25 °C (+2.3 °C above the long-term average, 2014–2021) and precipitation of 64.05 mm (−164 mm below the long-term average, 2014–2021) in July and August (Figure S1), which broke the high temperature and low precipitation records during the same period recorded by the Xujiahui Meteorological Station in Shanghai since 1873. Specifically, the occurrence of a “drought heatwave” was accompanied by high VPD and low volumetric soil water content, with a severe lack of precipitation in July and August 2022 (Figure 2b,c). This created an unpredictable extreme situation for the tree species, with obvious drought and water shortage characteristics during the critical growth season for trees, which should immediately impact and cause an instantaneous response of both large-scale vegetation index and physiological traits.

Figure 2. Annual monthly meteorological data statistics of the study area in 2022. (a) The red curve represents the monthly mean temperature (T), and the gray bar graph represents the monthly soil temperature (T_s). (b) The red curve represents the relative air humidity (RH), and the gray curve represents the soil moisture (M_s). (c) Water vapor pressure deficit (VPD). (d) The volumetric water content of the soil (VWC). (e) Precipitation (Ppt). The gray rectangular area is used to highlight the period of the drought–heatwave event. All abbreviations are shown in

Table 1. Summary of the abbreviations used in this study.

	Abbreviation	Full Name/Scientific Name	Unit
Climate variables	T	Air temperature	(°C)
	RH	Relative air humidity	(%)
	VPD	Vapor pressure deficit	(kPa)
	Ts	Soil temperature	(°C)
	Ms	Soil moisture content	(%)
	VWC	Volumetric water content	(m3/m3)
	Ppt	Precipitation	(mm)
Leaf traits	RWC	Leaf relative water content	(%)
	SPAD value	Relative value of chlorophyll content	-
	gs	Stomatal conductance	(mol−1 m−2 s−1)
Hydraulic traits	Ks	Stem hydraulic conductivity	(kg m s−1 MPa−1)
	Ψmidday	Midday water potential	(MPa)
	PLC	Percent loss of hydraulic conductivity	(%)
	Safety margin	The difference between water potential and P50	(MPa)
	P50	Water potential at 50% ofmaximum hydraulic conductivity	(MPa)
Remote sensing	NDVI	Normalized difference vegetation index	-
Species	KP	Koelreuteria paniculata Laxm.	-
	CC	Cinnamomum camphora (L.) Presl	-
	MG	Metasequoia glyptostroboides Hu and W. C. Cheng	-
	TD	Taxodium distichum var. imbricatum (Nuttall) Croom	-

.

Figure 2. Annual monthly meteorological data statistics of the study area in 2022. (a) The red curve represents the monthly mean temperature (T), and the gray bar graph represents the monthly soil temperature (T_s). (b) The red curve represents the relative air humidity (RH), and the gray curve represents the soil moisture (M_s). (c) Water vapor pressure deficit (VPD). (d) The volumetric water content of the soil (VWC). (e) Precipitation (Ppt). The gray rectangular area is used to highlight the period of the drought–heatwave event. All abbreviations are shown in Table 1.

3.2. Remote Sensing Data and Temperature Changes Are Closely Related

Comparing the daily temperature changes from May to September with the changes in the average NDVI values obtained via remote sensing monitoring of the four study sites, the NDVI values of the samples corresponding to the period when the maximum temperature exceeded 35 °C showed extremely low levels. Before the arrival of the hot drought event, the NDVI values of the sample sites also showed a gradual increase with the normal increase in temperature. When the extreme hot drought event was over, the NDVI of the study sites gradually decreased. In addition, under the same temperature conditions, there were obvious differences in the changes in the four study sites: the NDVI in the Liantang research site had the largest overall fluctuation—the difference between the maximum and minimum was the largest—while the changes in Dalianhu, Zhangma Village, and Nanyuewei were relatively small; among them, the changes in Dalianhu and Zhangma Village were similar (Figure 3).

3.3. Instantaneous Changes in Physiological Traits of the Four Tree Species during and after the Drought–Heatwave Event

Among the four dominant tree species of subtropical Eastern China, MG showed the greatest degree of change before and after the extreme drought and high-temperature event in terms of K_s, Ψ_midday, g_s, and the SPAD value (Figure 4b₃–e₃), while CC showed the lowest degree of change (Figure 4b₂–e₂). For the two major types of tree species in our study, i.e., broadleaf and coniferous tree species, the two broadleaf tree species and the two coniferous tree species showed obvious overall differences in terms of the water potential, SPAD value, and g_s: the fluctuations in the broadleaf tree species were smaller overall than those of the coniferous trees during the hot drought event (Figure 4), showing that broadleaf tree species are more resistant to extreme heat and drought events than coniferous tree species. However, just focusing on the physiological traits of these two types of tree species in the subsequent period following the extreme high temperature and drought event, it can be seen that, compared with the small increase in the two broadleaf forest species, the overall increase in the two coniferous forest species was larger, showing a stronger recovery ability than the broadleaf tree species (Figure 4).

The extreme heat and drought event caused the Ψ_midday of the four tree species to decrease significantly (Figure 5c). Among them, the safety margin in MG and TD fell to a negative value during the drought–heatwave event, and the rate of change was larger than that of KP and CC (Figure 5b), which also confirms the stronger resistance of broadleaf tree species in coping with hot drought events and the stronger resilience of coniferous tree species. In addition, the variation in the PLC level of these four tree species showed a significant increase during the hot drought event (Figure 5a), accompanied by a decrease in K_s (Figure 4b).

3.4. Relationship between Physiological Characteristics and Meteorological Factors

The variation in the meteorological factors (including T, VPD, RH, M_s, VWC, and T_s) before and after the hot drought event and with the changes in the physiological characteristics of trees were analyzed to obtain the correlation relationships. Specifically, there was a negative correlation between T, VPD, T_s, and various physiological traits, while RH, M_s, and VWC showed a positive correlation with the physiological indicators (Figure 6). Among them, the correlation between K_s and RH was the strongest, and the correlation with M_s was the weakest. The correlation between Ψ_midday and M_s was the strongest, and the correlation with RH was the weakest. The SPAD value showed a strong correlation with various environmental factors, showing a positive correlation with RH and VWC, and a negative correlation with T, T_s, and VPD (Figure 6).

4. Discussion

This study aimed to assess the transient in situ physiological integrity and large-scale vegetation indices from satellite image data of four dominant subtropical tree species, in particular with respect to hydraulic conductivity, xylem embolism, stomatal conductance, and the NDVI, during the severe 2022 summer drought–heatwave event in the subtropical region of China. For this purpose, in four dominant tree species in Shanghai, Eastern China, we recorded the instantaneous changes in physiological traits, such as hydraulic conductivity and leaf gas exchange, in the field and the NDVI changes from remote sensing at the same time and site in three stages (before, during, and after the extreme drought–heatwave climate event).

4.1. Response of NDVI Variation from Remote Sensing to Drought and Heatwave

The forests growing in the four sites we studied (Figure 1) experienced an extreme compound high-temperature–drought climate event during summer 2022 (Figure 2), and our data indicate that the hot drought event in summer 2022 caused a sudden drop in the NDVI in all four forest sites (Figure 3). At the same time, we also observed a fast recovery trend of the NDVI in the four sites after the hot drought event (Figure 4a₁–a₄), with the NDVI returning to over 70% of its previous level in these four forests. Similar results were reported by S. Haberstroh et al., who monitored a significant decrease in the NDVI in trees in the year of a high-temperature–drought event and a gradual return to normal levels in the following year [66]. This further suggests that the NDVI does play an important role in characterizing tree responses to drought events [67,68]. Compared to these studies, the analysis using remote sensing images from the Landsat series of satellites allowed us to monitor the same sample twice in one month, rendering the amplitude of NDVI change in a short period of time and the instantaneous response more intensive.

The small area of monitoring also makes the NDVI of a forest plot highly susceptible to extreme exceptions in the forest area. In addition, the effects of this “thermal drought” of extreme combined dry–hot climatic events [13] greatly threaten tree water availability [14], resulting in extremely low levels of tree moisture status for a short period of time. This will undoubtedly lead to the discoloration of leaves in a short period of time [32], premature abscission of leaves [33], and even withering, resulting in a sharp decline in the NDVI of the trees in a short period of time. Therefore, we infer that such a large NDVI anomaly is related to leaf wilt in some trees in the plot.

4.2. Responses of Hydraulics and Photosynthesis in the Field

In this study, we found that when the extreme drought–heatwave event occurred, the K_s, Ψ_midday, g_s, RWC, and SPAD values showed an immediate decline in the four tree species. In terms of the correlation between the physiological characteristics of the four tree species and various environmental factors, an intense rise in T, T_s, and VPD will indirectly lead to drought by affecting RH, M_s, and VWC, thereby instantly decreasing the water availability in the soil and air. In fact, heatwave events often coincide with droughts under global climate change [28].

Drought can exacerbate water absorption difficulties [69] and water dispersion loss in tree roots by affecting the soil moisture content and air moisture content [70]. A high-temperature event will increase the VPD in the air [71,72] and the leaf temperature [73], as well as expand the stomatal conductance [74,75,76], thereby aggravating the transpiration of trees and making the water deficit even worse [77,78]. Drought-induced embolism blocks the water transportation within the xylem of trees, resulting in a decrease in hydraulic conductivity [79] and further destruction of the hydraulic system integrity with “hydraulic failure”. High-temperature events could also destroy chlorophyll, reduce enzyme activity, disrupt many intermediate reactions, lead to the cessation of photosynthesis [80], and affect the growth and development of trees. Therefore, the indicators K_s, Ψ_midday, g_s, and RWC are direct embodiments of the influence of dry heat environments on the internal hydraulic balance state of trees, and the SPAD value characterizes the photosynthetic state of trees during hot drought events.

4.3. Relationship between Hydraulic Functioning in the Field and NDVI

The conjunction of drought and heatwaves has been increasingly considered to be the primary trigger of worldwide tree dieback and mortality in large-scale monitoring [81,82]. Drought- and heatwave-induced canopy damage observed on individual trees manifests as severe leaf scorch and shedding, which is related to the coupled hydraulic and thermal defects; this may also be related to the simultaneous decline in the NDVI during drought and heatwave events. Hot droughts with a higher VPD and limited available water content cause a further increase in xylem tension with a decline in the water potential, accompanied by drought-induced xylem embolism spreading; therefore, xylem hydraulic conductivity decreases since the number of embolized vessels increases, hence provoking hydraulic dysfunction [83,84]. Destruction of trees’ hydraulic system could lead to distal twigs, leaf dehydration, and thus severe leaf scorch and shedding under dry–hot stress [9,85], which might ultimately give rise to a decrease in the NDVI in remote sensing images. In our results, a rapidly decreasing K_s detected through xylem embolism with an increasing PLC was clearly shown during the dry–hot event (Figure 4 and Figure 5), which correlated with the leaf scorch and shedding on individual trees exhibiting the phenomenon of serious canopy damage on a large-scale following the drought–heatwave stress condition (

Table 2. The dieback percentage of four major tree species distributed over four research sites during the drought–heatwave event.

Tree Species	Distributed Research Field(s)	Dieback Percentage
Koelreuteria paniculata Laxm.	Dalianhu	60%
Cinnamomum camphora (L.) Presl	Nanyuewei/Dalinhu	40%
Metasequoia glyptostroboides Hu and W. C. Cheng	Liantang/Zhangma Village	50%
Taxodium distichum var. imbricatum (Nuttall) Croom	Liantang/Zhangma Village	40%

). In terms of the associated intensity, temperature and drought factors had the greatest effect on xylem hydraulic conductivity in the leaves (Figure 6), accompanied by tree canopy changes, as observed in the satellite images [86]. Under the drought–heatwave event, physiological processes and the vegetation index obtained via remote sensing both showed a quick response, with the hydraulic function responding on the scale of minutes to days, and the vegetation canopy changes occurring on the scale of days.

4.4. Trade-Off between Resistance and Resilience of Hydraulic Traits to Drought–Heatwave

Our results suggested that there may be a trade-off between the instantaneous resistance and resilience of hydraulic conductivity and stomatal conductance to drought and heatwaves among these four subtropical tree species in Eastern China, as the species with a lower degree of decline in K_s and g_s during the hot drought had a shorter recovery after the stress conditions. In general, the degree of decline in the hydraulic traits of the broadleaf species was lower than that of the coniferous species during the high-temperature–drought event (Figure 4), which was related to the higher Ψ_midday and larger hydraulic safety margin of the broadleaf species (Figure 5), reflecting the stronger resistance of broadleaf species to dry–hot climate events compared with coniferous forest species in the study area. After the drought–heatwave event, the overall recovery of K_s and g_s in the coniferous forest species was greater than that of the broadleaf forest species (Figure 4), reflecting the relatively strong resilience of coniferous forest species compared to broadleaf forest species after extreme hot drought climate events. Indeed, in previous research, MG and TD were found to be less drought-tolerant and preferred to grow under humid conditions [87]; therefore, they might have lower drought resistance compared with KP and CC under the same hot drought stress.

Specifically, the two fir trees showed a clear difference in sensitivity and tolerance to hot drought, with different changes in K_s and Ψ_midday under the stress conditions. Among these two coniferous forest species, in the face of the extreme compound drought–heatwave climate event, MG was more sensitive, while TD showed tolerance characteristics. This is mainly reflected in the fact that when MG experienced the extreme climate event, the relevant hydraulic traits showed a greater decrease than those of TD (Figure 4a₃,a₄,b₃,b₄). The study of LI Shubin et al. pointed out that such differences in characteristics are better reflected in the different hydraulic transportation supply strategies of different fir trees, with tolerant firs mainly adopting proactive survival strategies to respond to drought adversity, such as increasing the total diameter of the main stem and the cross-sectional area of the sapwood to improve their water channeling capacity and reducing the diameter of their side branches to reduce water consumption. Sensitive fir trees adopt passive tolerance strategies, such as reducing the diameter of the main stem to reduce water conductivity and increasing the Huber value of the side branches to bear less leaf mass. Furthermore, whether fir trees can carry out a structural instantaneous response in the face of such sudden weather events needs to be further verified [88].

In addition, KP had a larger hydraulic safety margin (Figure 5a) compared to the relatively smaller resilience and safety margin of CC among the broadleaf species, with a higher degree of recovery for various indicators (especially K_s and the SPAD value) after the extreme climatic event (Figure 4a,c), which we speculate may be related to the survival strategy of KP, which, being a deciduous tree species, seizes opportunities to grow quickly during the growing season. KP adopts a “risk-taking” growth strategy, having a strong ability to regulate physiological states in a short period of time, while CC grows evergreen all year round and exhibits a more “conservative” growth strategy. In order to maintain a relatively stable growth state, CC will adopt relatively conservative strategies for growth and resistance to extreme weather, which leads to its inability to quickly adjust its physiological state to recover from previous damage in the face of extreme compound dry–hot climate events.

5. Conclusions

In summary, we found that the drought–heatwave event of summer 2022 caused an instantaneous response in both the large-scale vegetation index obtained via remote sensing data and water-related physiological traits based on in situ measurements in the field for four dominant subtropical afforestation tree species in Shanghai, Eastern China. Hydraulic and photosynthetic processes were synchronized in responding to the hot drought event, with a quickly reduced xylem hydraulic capacity, leaf gas exchange, and water potential, indicating that there is coordination between them. The exacerbated hydraulic system dysfunction during hot drought stress leading to increased tree crown dieback may be associated with the decrease in the NDVI obtained from the satellite images. Furthermore, a trade-off between resistance and resilience in hot drought-induced hydraulic dysfunction was found among the four subtropical tree species. It is important to note that our study was based on the fact that the forest is a single species forest, which is less applicable when comparing species if it is a mixed forest. Accurately matching remote sensing images on a large scale with field surveys on a small scale in temporal and spatial terms should be improved in future research.