Preceding Phenological Events Rather than Climate Drive the Variations in Fruiting Phenology in the Desert Shrub Nitraria tangutorum

Fruit setting and ripening are crucial in the reproductive cycle of many desert plant species, but their response to precipitation changes is still unclear. To clarify the response patterns, a long-term in situ water addition experiment with five treatments, namely natural precipitation (control) plus an extra 25%, 50%, 75%, and 100% of the local mean annual precipitation (145 mm), was conducted in a temperate desert in northwestern China. A whole series of fruiting events including the onset, peak, and end of fruit setting and the onset, peak, and end of fruit ripening of a locally dominant shrub, Nitraria tangutorum, were observed from 2012 to 2018. The results show that (1) water addition treatments had no significant effects on all six fruiting events in almost all years, and the occurrence time of almost all fruiting events remained relatively stable compared with leaf phenology and flowering phenology after the water addition treatments; (2) the occurrence times of all fruiting events were not correlated to the amounts of water added in the treatments; (3) there are significant inter-annual variations in each fruiting event. However, neither temperature nor precipitation play key roles, but the preceding flowering events drive their inter-annual variation.

Nitraria tangutorum is a pioneer species widely distributed in the northwestern regions of China. It has a high tolerance to drought, high temperature, cold, salinity, alkalinity, bareness, wind erosion, and sand burial and plays a pivotal role in the desert ecology by fixing sands as well as slowing down deserts' expansion. The fruit of N. tangutorum, called "desert cherry" (Figure 1C), is a native folk medicine used to treat dyspepsia, neurasthenia, abnormal menstruation, and heart disease in western China [20][21][22]. Furthermore, the berries provide Eremias multiocellata and Eremias przewalskii with the water they need during neurasthenia, abnormal menstruation, and heart disease in western China [20][21][22]. Fu thermore, the berries provide Eremias multiocellata and Eremias przewalskii with the wate they need during the summer heat [23]. Moreover, another valuable medicine Cynomorium songaricum Rupr., which is parasitic on the roots of N. tangutorum, has a rele vant socioeconomic importance that determines the management of N. tangutorum. Pre cipitation at the present study area has shown an increasing trend over the last 40 year (Figure 2A), matching a regional trend in northwestern China identified in [24][25][26]. Th precipitation increase is causing the advance of leaf unfolding, the postponement of lea cessation, and the prolongation of the growing season in N. tangutorum [27,28]. Howeve we know little about whether the fruiting phenology is altered. Characterizing the impact of precipitation increase on the fruiting phenology of N. tangutorum is crucial given it ecological and economical importance. Here, we conducted a seven-year in situ simulate water addition experiment to evaluate the response of fruiting phenology to precipitatio increase. Five water addition treatments were designed to simulate precipitation increase natural precipitation (Control) and natural precipitation plus an extra 25% (+25%), 50% (+50%), 75% (+75%), and 100% (+100%) of the local annual mean precipitation (145 mm 1978-2008). The whole series of the phenological events of N. tangutorum, including lea unfolding (onset and end), flowering (onset and end), fruit setting (onset, peak, and end fruit ripening (onset, peak, and end), and leaf cessation (leaf coloring and leaf fall), wa detected. The aims of this study were to address the following specific questions: (1) How will water increase affect the patterns of fruiting phenology? (2) Do fruiting events shi consistently with the other phenological events? (3) Does water drive the variation in frui ing time?   neurasthenia, abnormal menstruation, and heart disease in western China [20][21][22]. Fur thermore, the berries provide Eremias multiocellata and Eremias przewalskii with the wate they need during the summer heat [23]. Moreover, another valuable medicine Cynomorium songaricum Rupr., which is parasitic on the roots of N. tangutorum, has a rele vant socioeconomic importance that determines the management of N. tangutorum. Pre cipitation at the present study area has shown an increasing trend over the last 40 year (Figure 2A), matching a regional trend in northwestern China identified in [24][25][26]. Th precipitation increase is causing the advance of leaf unfolding, the postponement of lea cessation, and the prolongation of the growing season in N. tangutorum [27,28]. Howeve we know little about whether the fruiting phenology is altered. Characterizing the impact of precipitation increase on the fruiting phenology of N. tangutorum is crucial given it ecological and economical importance. Here, we conducted a seven-year in situ simulate water addition experiment to evaluate the response of fruiting phenology to precipitatio increase. Five water addition treatments were designed to simulate precipitation increase natural precipitation (Control) and natural precipitation plus an extra 25% (+25%), 50% (+50%), 75% (+75%), and 100% (+100%) of the local annual mean precipitation (145 mm 1978-2008). The whole series of the phenological events of N. tangutorum, including lea unfolding (onset and end), flowering (onset and end), fruit setting (onset, peak, and end fruit ripening (onset, peak, and end), and leaf cessation (leaf coloring and leaf fall), wa detected. The aims of this study were to address the following specific questions: (1) How will water increase affect the patterns of fruiting phenology? (2) Do fruiting events shi consistently with the other phenological events? (3) Does water drive the variation in frui ing time?

Inter-Annual Dynamics of Meteorological Factors
The climate at the study site is markedly seasonal, with a wet/warm season from May to September and a dry/cold season occurring from October to April. The annual precipitation at the study site showed an increasing pattern in the past 40 years ( Figure 1A). The annual mean temperature (AMT), winter mean temperature (T Win , December to February), spring mean temperature (T Spr , March to May), summer mean temperature (T Sum , June to August), and autumn mean temperature (T Aut , September to November) all showed no significant increasing or decreasing trends (Bao et al., 2020). The annual mean precipitation (AMP) in 2012,2013,2014,2015,2016,2017, and 2018 was 213.3, 59.1, 95.2, 147, 189.2, 86.0, and 59.1 mm, respectively. Winter rainfall was scarce (P Win , December to February), occurring only in three of the seven years, and the amounts were very low, with less than 5 mm (Bao et al., 2020). Spring (P Spr , March to May), summer (P Sum , June to August), and autumn (P Aut , September to November) rainfall varied from 2012 to 2018 with no increasing or decreasing patterns (Bao et al., 2020). This study focused on the variations in the accumulated precipitation and temperature prior to and during the fruiting period (1 January-31 August). The accumulated precipitation in 2012, 2013, 2014, 2015, 2016, 2017, and 2018 was 192.3, 51.5, 51.4, 80.9, 160.5, 63.8, and 56.2 mm, respectively. More than 95% of the rain fell during the fruiting period (May to August). These amounts are very close to the annual precipitation. Based on the deviation from the mean value of the seven years, 2012 and 2016 were "above-average" (i.e., wet) years; 2013 was an "ultra below-average" (i.e., extremely dry) year; 2014, 2017, and 2018 were "below-average" (i.e., dry) years; and 2015 was an "average" year ( Figure 2B).

Changes in Fruit Setting
The linear mixed model analysis showed that the water addition treatments had no significant effects on the onset (OFS), peak (PFS), and end time (EFS) of fruit setting (all p > 0.05, Table 1) over the seven years (2012)(2013)(2014)(2015)(2016)(2017)(2018). The year affected all three events significantly (all p < 0.05, Table 1) over the seven years, and there was no interaction between water addition treatment and year (all p > 0.05, Table 1). Table 1. The p-values of the linear mixed model (MIXMOD) analysis on the fixed effects of water addition treatments (water), year, and their interaction on the fruiting events and the duration of the fruiting stage (DF) from 2012 to 2018. OFS, PFS, and EFS represent the onset, peak, and end of the fruit setting stage, respectively. OFR, PFR, and EFR represent the onset, peak, and end of the fruit ripening stage, respectively. The shifting direction of fruit setting events was not consistent among water addition treatments and among years in the 2012-2018 period ( Figure 3).

OFS
The occurrence time of OFS was advanced in three (2012, 2013, and 2015) and delayed in one (2014) of the seven years under all four water addition treatments (+25%, +50%, +75%, and +100%) (all p > 0.05 with the exception of the +50% water addition treatment in 2015, Figure 3A). The shifting directions were not consistent among water addition treatments in the other years ( Figure 3A). On average, over the seven years (2012-2018), the occurrence of OFS in the +25%, +50%, +75%, and +100% water addition treatment plots was not significantly changed compared with the control (all p > 0.05, Figure 3A Figure 3B). The shifting directions were not consistent among water addition treatments in the other years ( Figure 3B). On average, over the seven years (2012-2018), the occurrence of PFS in +25%, +50%, +75%, and +100% water addition treatment plots were not significantly changed compared with control (all p > 0.05, Figure 3B, last column).
The occurrence time of EFS was advanced in two (2012 and 2018) and delayed in one (2014) of the seven years under all four water addition treatments over the seven years (2012-2018) (all p > 0.05 with the exception of the +50% and +100% water addition treatments in 2012, Figure 3C). The shifting directions were not consistent among water addition treatments in the other years ( Figure 3B). On average, over the seven years (2012-2018), the occurrence of EFS in the +25%, +50%, +75%, and +100% water addition treatment plots was not significantly changed compared with the control (all p > 0.05, Figure 3C, last column). (all p > 0.05, Figure 3B). The shifting directions were not consistent among water addition treatments in the other years ( Figure 3B). On average, over the seven years (2012-2018), the occurrence of PFS in +25%, +50%, +75%, and +100% water addition treatment plots were not significantly changed compared with control (all p > 0.05, Figure 3B, last column). The occurrence time of EFS was advanced in two (2012 and 2018) and delayed in one (2014) of the seven years under all four water addition treatments over the seven years (2012-2018) (all p > 0.05 with the exception of the +50% and +100% water addition treatments in 2012, Figure 3C). The shifting directions were not consistent among water addition treatments in the other years ( Figure 3B). On average, over the seven years (2012-2018), the occurrence of EFS in the +25%, +50%, +75%, and +100% water addition treatment plots was not significantly changed compared with the control (all p > 0.05, Figure 3C, last column).

Changes in Fruit Ripening
The linear mixed model analysis showed that the water addition treatments had no significant effects on the onset (OFR), peak (PFR), and end time (EFR) of fruit ripening (all p > 0.05, Table 1) over the seven years (2012)(2013)(2014)(2015)(2016)(2017)(2018). The year affected all three events significantly (all p < 0.05, Table 1) over the seven years, and there was no interaction between water addition treatment and year (all p > 0.05, Table 1).
The shifting directions of fruit ripening events were not consistent among water addition treatments and among years over 2012-2018 ( Figure 4).

Changes in the Duration of Fruiting Period
The linear mixed model analysis showed that the water addition treatments had no significant effects on the duration of the fruiting period (DF) (p > 0.05, Table 1) over the seven years (2012-2018). The year had a significant effect on it (p < 0.05, Table 1). There was no interaction effect between water addition treatment and year (p > 0.05, Table 1).

+25%
+50% +75% +100% There were no significant effects of water addition treatments on the occurrence time of OFR in six of the seven years, except 2012 (all p > 0.05, Figure 4A). On average, in the seven years (2012-2018), the occurrence of OFR in the +25%, +50%, +75%, and +100% water addition treatment plots was not significantly changed compared with the control (all p > 0.05, Figure 4, last column).

Δdays
The occurrence time of PFR was advanced in 2012 under all four water addition treatments (all p < 0.05, Figure 4B). However, it was delayed in 2016 under all four water addition treatments (all p < 0.05, Figure 4B). There were no significant effects of water addition treatments on the occurrence time of PFR in the other five years from 2012 to 2018 (all p > 0.05, Figure 4B). On average, over the seven years (2012-2018), the occurrence of OFR in the +25%, +50%, +75%, and +100% water addition treatment plots was not significantly changed compared with the control (all p > 0.05, Figure 4B, last column).
The occurrence time of EFR was delayed in six of the seven years under all four water addition treatments, with the exception in 2012 (all p > 0.05, Figure 4C). On average, over the seven years (2012-2018), the occurrence time of EFR in the +25%, +50%, +75%, and +100% water addition treatment plots was not significantly changed compared with the control (all p > 0.05, Figure 4, last column).

Changes in the Duration of Fruiting Period
The linear mixed model analysis showed that the water addition treatments had no significant effects on the duration of the fruiting period (DF) (p > 0.05, Table 1) over the seven years (2012-2018). The year had a significant effect on it (p < 0.05, Table 1). There was no interaction effect between water addition treatment and year (p > 0.05, Table 1).

Comparison with Other Phenological Events
On average, over the seven years (2012-2018), the water addition treatments significantly advanced the leaf unfolding and flowering events and delayed the leaf cessation events. For example, the end of leaf unfolding time (ELF) in the +25, +50, +75, and +100% water addition treatments was advanced by 5.07, 9.29, 9.29, and 7.86 d, respectively. The end of leaf coloring time (ELC) in the +25, +50, +75, and +100% water addition treatments

Comparison with Other Phenological Events
On average, over the seven years (2012-2018), the water addition treatments significantly advanced the leaf unfolding and flowering events and delayed the leaf cessation events. For example, the end of leaf unfolding time (ELF) in the +25, +50, +75, and +100% water addition treatments was advanced by 5.07, 9.29, 9.29, and 7.86 d, respectively. The end of leaf coloring time (ELC) in the +25, +50, +75, and +100% water addition treatments was delayed by 2.32, 6.00, 9.07, and 12.32 d, respectively. Compared with leaf phenology, the variation in flower phenology was lower. The OFL in the +25, +50, +75, and +100% water addition treatments was significantly advanced by 0.43, 1.29, 2.43, 1.14 d, respectively (all p < 0.05). The EFL in the +25%, +50%, +75%, and +100% water addition treatments was significantly advanced by 1.89, 1.46, 2.25, and 1.14 d, respectively (all p < 0.05). By contrast, the changes in fruiting events were even smaller compared with the flowering events. We found that the responses of all fruiting events to the water addition treatments were relatively stable compared to the leaf and flowering phenology ( Figure 5). The effects of the water addition treatments on fruiting events, namely OFS, PFS, EFS, OFR, PFR, and EFR, were not significant ( Figure 6, all p > 0.05).

Figure 5.
Relative changes (△days; mean ± SE) in the duration of fruiting period (DF) of Nitraria tangutorum after water addition treatments (+25%, +50%, +75%, +100%) compared with the control. Positive and negative values represent shortened and prolonged days, respectively. Avg. represents average values over the period from 2012 to 2018. * Significant differences from control.

Comparison with Other Phenological Events
On average, over the seven years (2012-2018), the water addition treatments significantly advanced the leaf unfolding and flowering events and delayed the leaf cessation events. For example, the end of leaf unfolding time (ELF) in the +25, +50, +75, and +100% water addition treatments was advanced by 5.07, 9.29, 9.29, and 7.86 d, respectively. The end of leaf coloring time (ELC) in the +25, +50, +75, and +100% water addition treatments was delayed by 2.32, 6.00, 9.07, and 12.32 d, respectively. Compared with leaf phenology, the variation in flower phenology was lower. The OFL in the +25, +50, +75, and +100% water addition treatments was significantly advanced by 0.43, 1.29, 2.43, 1.14 d, respectively (all p < 0.05). The EFL in the +25%, +50%, +75%, and +100% water addition treatments was significantly advanced by 1.89, 1.46, 2.25, and 1.14 d, respectively (all p < 0.05). By contrast, the changes in fruiting events were even smaller compared with the flowering events. We found that the responses of all fruiting events to the water addition treatments were relatively stable compared to the leaf and flowering phenology ( Figure 5). The effects of the water addition treatments on fruiting events, namely OFS, PFS, EFS, OFR, PFR, and EFR, were not significant ( Figure 6, all p > 0.05).     Figure 7). For the individual fruiting event OFS, which was obviously affected by water addition treatments in three of the seven years, its response patterns to the w addition treatments among the three years were not consistent. A positive quadratic r tionship in 2014 and negative quadratic relationships in 2012 and 2015 were determi ( Figure 7A).
In Tables 2 and 3, AMP and AMT represent the annual mean precipitation and annual mean temperature, respectively; T Win , T Spr , T Sum , and T Aut represent the mean temperature during winter (December to February), spring (March to May), summer (June to August), and autumn (September to November), respectively; P Win , P Spr , P Sum , and P Aut represent the accumulated precipitation during winter (December to February), spring (March to May), summer (June to August), and autumn (September to November), respectively; and OFS, PFS, and EFS represent the onset, peak, and end of the fruit setting period, respectively.  All p < 0.05 except +100% All p < 0.05 All p < 0.05 All p < 0.05 All p < 0.05 except +25% Ctrl and +100%: p < 0.05 Ctrl, +25%, and +100%: p < 0.05

Effects of Water Addition Treatments on Fruiting Events
Plants that live in deserts have to contend with stressful conditions characterized by sandy soils with low moisture and low fertility. These species are aridity tolerators and are also water-sensitive. The leaf unfolding and leaf coloring events were advanced or delayed synchronously with water addition amount increases [27]. The results do not mean that the fruiting phenology will follow the same pattern. In fact, this study found that the fruit setting and ripening time and the duration of the fruiting period were not significantly different from the control in most cases under the different water addition treatments (+25%, +50%, +75%, and +100%) and in most years (Figures 3-5). The fruiting phenology remained relatively stable relative to the significant changes in leaf unfolding, leaf cessation, and even flowering after all four water addition treatments (+25%, +50%, +75%, and +100%). Similarly, rainfall was not significantly correlated with the proportion of flowers setting fruits in the Andean shrub Befaria resinosa (Ericaceae) [29]. A double precipitation treatment did not significantly affect the fruiting phenology in a tallgrass prairie in the south-central Great Plains in the United States [14]. A long-term change in fruiting phenology in a West African lowland tropical rainforest was also not explained by rainfall [30]. The exception of the significant advance in fruit ripening (OFR, PFR, and EFR) in 2012 ( Figure 4) might have been caused by two possible reasons. First, it may relate to the extreme wet situation during the period from EFS to PFR (26.9 mm on day 178 and 32.1 mm on day 185) in 2012. The effect of water stress caused by extra water addition over a relatively extremely wet season may have resulted in the early end of fruit ripening [28]. Second, the life form can modulate the effects of precipitation on fruiting phenology [10]. In this study, the higher-magnitude advances in EFS in 2012 after water addition treatments relative to the other six years ( Figure 3C) caused the significant advancing of the successional fruiting phenology. Similar results have also been reported by other authors. Accordingly, the significant advances in EFS after water addition treatments in 2012 ( Figure 3C) were caused by the preceding events. In this special, extremely wet year, all phenological events, including leaf, flowering, and fruiting events, were significantly advanced [28].

Relative Stable Fruiting Phenology Relative to Leaf and Flower Phenology
The slight variation in fruiting events after the water addition treatments and the relatively stable response pattern with the increasing water addition amounts (Figures 6 and 7) indicated that the fruiting phenology of N. tangutorum is independent from precipitation. Combined with the results showing that the inter-annual variation in fruiting events was not related to the temperature and precipitation factors (Tables 2 and 3), in this study, it is reasonable to attribute this stability to inherent properties. Firstly, plant phylogenetic constraints play a pivotal role in shaping the shifts in fruiting phenology [17,31,32]. Secondly, biotic interactions could be important drivers in the fruit setting and ripening [33]. Berries of N. tangutorum are an important component of many food webs. For example, Eremias multiocellata adjust a large proportion of their diet to berries of N. tangutorum during hot summers to make up for the loss of surface moisture [23]. Changes in berry setting and ripening may have significant consequences for trophic interactions and ecosystem function [34,35]. Long-term coevolution has played a major role in driving the niche conservatism among species in the harsh and variable desert ecosystem. For example, in Sahelian shrubs, morphological and physiological adaptations strongly contribute to the relative independence of their activity from water availability [36]. These results indicate a strong plasticity in the fruiting phenology of N. tangutorum in response to precipitation increase.

Drivers of Inter-Annual Variations in Fruiting Phenology
The timing of fruiting phenology is the result of internal and external cues [37]. Abiotic environmental conditions such as rain [38][39][40][41][42], temperature [5,14,15,43], photoperiod, and ENSO [41,44] have been shown to play a significant role in the timing of fruiting phenology. Meanwhile, no significant correlation was determined between precipitation (winter, spring, summer, autumn, and annual) and temperature (winter, spring, summer, autumn, and annual) factors and any of the six fruiting events in this study (all p > 0.05, Tables 2 and 3). We found that the inter-annual changes in fruit setting and ripening were significantly correlated with the changes in the preceding events OFL and EFL (almost all p < 0.05, Figure 8). In many cases, the phenological patterns have been regulated by natural selection [45]. Obviously, the strength of phylogenetic conservatism outweighed the climate effects on the fruiting phenology in this study. Taken together, our results in this study suggest that the fruiting time of N. tangutorum is dependent on the preceding flowering events rather than precipitation or temperature. Similar shifts have been reported in other ecosystems. For example, in one study, temporal variation in the flowering time explained 36% of the variation in the start of fruiting time [46]. The fruit setting patterns of baobabs, Adansonia digitata, were not significantly different between two years with a large rain fall difference in southern Africa [47]. A relatively stable timing of fruit setting and ripening could be associated with the formation of seeds satiating seed predators [1], which is essential for maintaining many aspects of biodiversity in a harsh and unpredictable desert environment [48].

Sample Area Description
Details of the experiment site, water addition treatments' design, and the monitoring of soil moisture and soil chemical properties were reported by Bao et al. (2020Bao et al. ( , 2021. In brief, the experiment was conducted at the Desert Ecosystem Water Addition Platform

Simulated Enhancement in Precipitation
Five water addition treatments were designed to simulate precipitation increases of 0% (Control), 25% (+25%), 50% (+50%), 75% (+75%), and 100% (+100%) of the local annual average precipitation (145 mm). The water addition treatments were applied equally every month from May to September. The additional water amounts were 0, 7.3, 14.5, 21.8, and 29.0 mm each time for the five water addition treatments, respectively. The water was pumped from a well near the plots into a water tank with water meters and then transported to each sprinkler. The sprinklers, with two automatically rotating spraying arms (6 m in length) that could uniformly sprinkle water over the treatment area, were installed on the top of each nebkha (plot). More detailed information on the experimental design and the irrigation system can be found in our previous publications [27,28].

Phenological Observations
The phenology observations were carried out from 2012 to 2018 following the standard protocols of Phenological Observation Methodology in China [49]. Phenology recording was conducted from March 2012 to November 2018. Phenological events for all shrubs in each nebkha (plot) were recorded every other day. A series of phenological events, including leaf unfolding (onset and end), flowering (onset and end), fruit setting (onset, peak and end), fruit ripening (onset, peak, and end), and leaf cessation (leaf coloring and fall), were recorded ( Figure 1, Table 4). Precipitation, air temperature, relative humidity, and evaporation data were recorded by a standard meteorological station near the experimental plots. The soil gravimetric water content (SWC) of the 0-30-cm soil layer was measured using the oven-drying method. The SWC significantly changed after the water addition treatments and was significantly linearly correlated to the water addition amounts. The influence degree of water addition on SWC was highest in spring, followed by autumn and then summer.

Leaf unfolding
Onset (OLU) At least one young leaf has completely extended and spread out completely from one or more leaf buds observed on the whole nebkha.
End (ELU) More than 90% of the young leaves on leaf buds of the whole nebkha have been completely spread.

Flowering
Onset (OFL) At least one flower bud on the whole nebkha has fully opened.
End (EFL) More than 90% of the flower buds in each nebkha have fully opened.

Fruit setting
Onset (OFS) At least one flower on the whole nebkha has at least one green berry visible.

Peak (PFS)
More than 50% of flowers on the whole nebkha have their green berries developed.

End (EFS)
More than 90% of flowers on the whole nebkha have their green berries developed.

Onset (OFR)
At least one berry on the whole nebkha is developing a red color.

Peak (PFR)
More than 50% of berries on the whole nebkha are developing red colors.

End (EFR)
More than 90% of berries on the whole nebkha are developing red colors.

Leaf cessation
End of leaf coloring (ELC) More than 90% of the leaves on the whole nebkha have turned yellow.
End of leaf fall (ELF) More than 90% of the leaves on the whole nebkha have fallen off.

Data Processing
The observed dates of fruiting events were first transformed to day of year. The duration of the fruiting period was calculated as the difference between the EFR and OFS. The relative change of the days (∆days) was used to test the effects of the water addition treatments on each event.
where day treat represent the day of a given event or the duration of the fruiting period in water addition plots, day ctrl represents the corresponding day in control plots, and n represents the number of the experimental years. Here, ∆days < 0 means that fruiting events (duration of fruiting period) were advanced (shortened) under water addition, while ∆days > 0 means that fruiting events (duration of fruiting period) were delayed (prolonged) under water addition.

Statistical Analysis
Simple linear regression analysis was used to determine the inter-annual trends of meteorological factors and fruiting events. Pearson correlation was used to analyze the relationships between fruiting events and meteorological factors. Linear mixed models were used to examine the effects of water addition treatments, year, and their interaction on fruiting events over the seven years (2012-2018). Water amount and year were used as fixed factors, while plot was used as a random factor. The dependent factor was the timing of different phenological events (Type I Sum of Squares was used). Duncan post hoc tests were used to determine pairwise differences for significant effects. One-way ANOVAs were used to test the effects of the water addition treatments on the timing of different fruiting events separately for each year. Homogeneity of variances was tested by Levine's tests. One-sample Kolmogorov-Smirnov tests were used to validate the normality of the data distribution. All statistical analyses were completed in SPSS 20.0 (IBM, Inc.; Armonk, NY, USA), and Microsoft Excel 2019 (Redmond, WA, USA) was adopted for plotting.

Conclusions
With increasing water amounts, the fruiting phenology of N. tangutorum maintained a relatively stable response compared to the leaf and flower phenology. The fruiting phenology was found to be independent from climatic factors, precipitation, and temperature in this study. The inter-annual variations in the fruiting phenology were not driven by water or temperature but by the preceding flowering events. However, responses of fruiting phenology to environmental changes are highly species-specific [10,12,14]. More studies across species are needed in the future to test the universality of the present results in a desert ecosystem.