Diverse impacts of day and night temperature on spring phenology in freshwater marshes of the Tibetan Plateau

Understanding the spatiotemporal variation in vegetation phenology of freshwater marshes and its response to climate change is significant to evaluate the global carbon cycling. Based on the climate and satellite‐derived vegetation index data, this work investigated the temporal and spatial variation of spring phenology and its response to climatic change in freshwater marshes of the Tibetan Plateau. We found that the start date of vegetation growing season (SOS) showed a significant advancing trend of 3.4 d decade−1 from 2001 to 2019 in freshwater marshes of the Tibetan Plateau. The increase of spring daytime and nighttime temperatures could advance the vegetation SOS over the marshes of the Tibetan Plateau. However, the effects of day maximum and night minimum temperature on freshwater marsh SOS were divergent at different altitudes: the advancing effect of warming night‐time and daytime temperature on the SOS was more significant in the regions below and above 4000 m, respectively.

The phenology of vegetation is an indicator of the change of vegetation and ecosystem function (Reed et al. 1994;Menzel 2002;White et al. 2003;Gu et al. 2009;Piao et al. 2011;Xia et al. 2015;Shen et al. 2022c). Wetlands are one of the three major ecosystems, which play a key part in global climate change and the carbon cycle (Middleton 2012;Shen et al. 2021c). At the present time, research on vegetation phenology mainly focuses on grassland and forest ecosystems Fu et al. 2020;Wang et al. 2020;Shen et al. 2022a), while research on wetland vegetation phenology is less common . Wetlands have unique environmental conditions which may lead to significant distinctions in the influence of climatic change on phenology compared with other ecosystems . Clarifying the impact of climatic change on wetland vegetation phenology will help promote the understanding of vegetation response to climatic change.
As a sensitive region of global climatic change, Tibetan Plateau has large areas of freshwater marsh wetlands Wang et al. 2021). The marsh vegetation of the Tibetan Plateau has important effects on the hydrological cycle, biogeochemical cycle, and carbon balance (Shen et al. 2020). So far, many studies have analyzed the spring phenological variations in vegetation and its response to climatic change over the Tibetan Plateau, but most of these studies focus on grassland ecosystems with few studies examining marsh ecosystems. Compared with grasslands, the freshwater marsh of the Tibetan Plateau is wetter, and the effects of climatic factors on vegetation phenology may be different. Thus, it is urgent to investigate the possible influences of climatic change on spring phenology of the Tibetan Plateau marsh vegetation.
Many studies found that the changes of grassland vegetation spring phenology were different at different elevations of the Tibetan Plateau (Piao et al. 2011;Shen et al. 2014). Until recently, however, whether the impacts of climatic change on vegetation spring phenology of the freshwater marsh of the Tibetan Plateau are distinct at different elevations is not clear. Based on the Normalized Difference Vegetation Index (NDVI) data, we investigated the temporal and spatial variation in marsh spring phenology of the Tibetan Plateau. Additionally, this study analyzed, for the first time, the possible influences of climatic change on freshwater marsh spring phenology at different elevations of the Tibetan Plateau.

Study region
The study area is situated in southwest China, and ranges from 26 N to 40 N and 73 E to 105 E (Fig. 1). The climate of the Tibetan Plateau is a plateau climate with low temperature and large diurnal temperature range. The annual precipitation in this region is about 400 mm (Li et al. 2021). The marsh types in the Tibetan Plateau include marsh meadows, herbaceous marshes, seasonal saltwater marshes, and inland salt marshes (Shen et al. 2021b).

Data
The climatic data are monthly minimum temperature, maximum temperature, mean temperature, and precipitation data of 106 meteorological stations over the Tibetan Plateau from 2001 to 2019. These climatic data were derived from the National Meteorological Information Center of China (http:// data.cma.cn/data). The NDVI data from 2001 to 2019 used in this paper is MOD13Q1 dataset provided by NASA (https:// ladsweb.modaps.eosdis.nasa.gov). The spatial and temporal resolutions of NDVI are 250 m and 16 d, respectively. Marshes were identified by using datasets on the distribution of freshwater marshes in China for 2000 and 2015 (Shen et al. 2021b).

Methods
This study adopted the ordinary Kriging method to interpolate the climate data of meteorological stations into the distribution of marshes in the Tibetan Plateau and resampled them to the same resolution as the NDVI data (Shen et al. 2021b). This study used the Polyfit-Maximum approach to obtain the start date of vegetation growing season (SOS) (Piao et al. 2006). In order to eliminate the impacts of covariates among multiple influencing factors (Shen et al. 2015), we used a temporal partial correlation analysis to investigate the correlations between the phenology of freshwater marsh vegetation and climate factors in spring (March-May) and previous winter (December-February), autumn (September-November), and summer (June-August). This study used the unchanged marsh (marsh in both two marsh maps) in the Tibetan Plateau as the research object to eliminate the impact of land cover change on the SOS results.
The data and metadata of this study are available in the Zenodo data repository (Shen et al. 2022b).

Results and discussion
Temporal and spatial variation in freshwater marsh spring SOS of the Tibetan plateau The SOS of freshwater marsh vegetation in the Tibetan Plateau was mainly between 120 and 160 DOY, and the average SOS was 139 DOY (19 May, or 18 May in leap year) ( Fig. 2A). This is mostly consistent with the finding of Ding et al. (2013) who found the SOS of Tibetan Plateau was 120-170 DOY. The annual average SOS gradually advances from the northwest to southeast of the Tibetan Plateau. The freshwater marsh vegetation turns green earlier at low elevations of the eastern Tibetan Plateau, but later at higher elevations in the central Tibetan Plateau (Fig. 2A), which is consistent with the finding of previous study (Piao et al. 2011). In highaltitude regions, the natural environment is cold and harsh ), which is not suitable for freshwater marsh vegetation growth.
From 2001 to 2019, the SOS of freshwater marsh vegetation significantly (p < 0.01) advanced in the Tibetan Plateau, with an advance trend of 3.4 d decade À1 (Fig. 3). The advance trend of the SOS obtained from this study was smaller than the calculated trend (6 d decade À1 ) by Ding et al. (2013), while similar to the result (4.6 d decade À1 ) of Tian (2015). This may be because that the remote sensing data used in this paper is MODIS, while Ding used SPOT data. Different remote sensing data and research objects could cause different results.

Correlations between the freshwater marsh vegetation SOS and climatic variables over the Tibetan Plateau
The partial correlation results indicated that the freshwater marsh vegetation SOS in the Tibetan Plateau was negatively correlated with precipitation, while the relationships were not significant (Table 1). In terms of temperature effect, there was a significant (p < 0.05) negative relationship of freshwater  marsh SOS with spring mean temperature over the Tibetan Plateau (Table 1). It indicates that spring temperature is a main factor influencing SOS, and the increasing spring temperature can advance the start date of freshwater marsh vegetation over the Tibetan Plateau. We found that the partial correlations between minimum and maximum temperatures and phenology were negative and asymmetric, with a stronger influence of minimum temperature on spring phenology than  that of maximum temperature (Table 1). It has been reported that the daytime maximum temperature can effectively complete the heat requirement for the unfolding of leaves, and is conducive to the carbon fixation of plants . Because heat accumulation mainly occurs in the daytime, the increasing maximum temperature will cause an advancing SOS in the Tibetan Plateau. In addition, due to high altitude and cold temperature, frost damage on ecosystem vegetation occurs frequently in the cold Tibetan Plateau. Therefore, warming nighttime minimum temperatures in the Tibetan Plateau can reduce the risk of frost damage, and relieve a restriction on plant developmental processes, thus promoting the breaking of ecodormancy (Körner, 2015;Shen et al. 2016). This could account for the larger sensitivity of freshwater marsh phenology to minimum temperature in the study area. Spatially, the positive relationships between the maximum temperature and SOS were only found at higher elevations in the southwest regions (Fig. 4A). For the effects of minimum temperature, we found that 81.84% of freshwater marsh vegetation SOS pixels were negatively correlated with spring minimum temperature. The most significant correlation was found in the eastern Tibetan Plateau (lower elevations). Partial relationships between the minimum temperature and SOS were positive for 18.16% freshwater marsh vegetation SOS pixels in the southwest Tibetan Plateau (higher elevations) (Fig. 4C). Shen et al. (2015) revealed that high preseason temperatures might reduce water availability through enhancing evapotranspiration and therefore delay the vegetation SOS in the arid regions of the northeast Tibetan Plateau. Different from the finding of Shen et al. (2015), we found that the increasing temperature could advance the green-up date in the eastern Tibetan Plateau. These differences may be due to the fact that the research object of Shen et al. (2015) was mainly arid or semi-arid grassland vegetation, while the research target of this study was freshwater marsh vegetation. As the freshwater marsh ecosystem has relatively sufficient water, the growth of freshwater marsh vegetation is not mainly limited by water availability (Zhang et al. 2017;Wang et al. 2021).
With the increase of altitude, in the regions below 4000 m, the effect of increasing spring minimum temperature on the advance SOS of freshwater marsh vegetation gradually weakened; while in the regions above 4000 m, the weakening change of this advancing effect tended to be gentle (Fig. 4D). The effects of nighttime minimum temperature on the spring phenology of freshwater marsh vegetation may be associated with two conditions for vegetation to green up. On the one hand, the increase of minimum temperature can reduce frost damage which promotes the advance of the green-up date of freshwater marsh vegetation. On the other hand, some chilling is needed to break the vegetation ecodormancy in the study area (Cong et al. 2017). In general, vegetation growing in cold regions needs a low-temperature threshold than in warm regions to break ecodormancy due to vegetation acclimation to colder temperature (Piao et al. 2011). Therefore, freshwater marsh vegetation has a higher demand for chilling in high-altitude (colder) regions, and the increase of minimum temperature will reduce chilling, thus weakening the advance influence on the freshwater marsh SOS. In the regions above 4000 m, with the increasing altitude and decreasing temperature, the ability to meet the demand for chilling gradually strengthened, so the influence of the minimum temperature on SOS gradually weakened. Because the increase of minimum temperature can reduce frost damage, the increase of the minimum temperature still has a weak advancing effect on the phenology.
With the increase of altitude, the advancing effect of increasing maximum temperature on the SOS gradually weakened in the regions below 4000 m (Fig. 4B). By contrast, the influence of increasing spring maximum temperature on the advance SOS of freshwater marsh vegetation gradually strengthened with the increase of altitude in the regions above 4000 m (Fig. 4B). The reason may be that in the regions below 4000 m, with the increase of altitude, the heat requirement for SOS declined in the colder areas (Yu et al. 2010). Therefore, the advancing effect of spring maximum temperature on freshwater marsh vegetation SOS gradually weakened with the increase of altitude in the Tibetan Plateau. In the regions above 4000 m, it is likely that the requirement of heat for vegetation green-up will not change significantly with the increase of altitude. However, the maximum temperature decreases sharply with the increase of altitude, which leads to gradual weakening of the ability to meet the accumulated temperature demand. As vegetation needs a certain thermal requirement to begin its spring growth, the advancing effect of warming spring maximum temperature on freshwater marsh vegetation SOS significantly strengthened with the increasing elevation in the regions above 4000 m of the Tibetan Plateau.
We found that the freshwater marsh vegetation SOS in the regions below 4000 m was mainly affected by the minimum temperature, while in the regions above 4000 m was mainly affected by the maximum temperature in the Tibetan Plateau. At lower altitude regions (below 4000 m), the heat accumulation during the daytime appears to meet the needs of vegetation growth. Therefore, the greater influence of minimum temperature on spring phenology could be because the increasing minimum temperature is conducive to reduce frost damage. In higher altitude regions (above 4000 m), due to cold environment, the ability to meet the heat requirement for spring green-up seems insufficient, thus the maximum temperature had a larger advancing influence on spring phenology.