Impact of diurnal unsymmetrical warming on soil respiration in an agroecological system of the Lhasa region

Purpose The impact of diurnal unsymmetrical rise in temperature on soil respiration (Rs) is not fully understood; thus, we explored such a warming influence on Rs in an agroecological system of the Lhasa. Materials and methods A field warming experiment (C: control; DW: daytime warming; NW: nighttime warming; DW+NW: daytime plus nighttime warming) was carried out in a naked barley ecological system. Results and discussion The DW, NW and DW+NW treatments dramatically increased soil temperature and decreased soil moisture but did not markedly modify Rs. The effects of DW and NW on soil respiration sensitivity (Q10) during the daytime and nighttime were different; they had no effects on daytime Q10 of Rs, but a significant inhibitory effect on nighttime Q10 of Rs. Conclusions A diurnal unsymmetrical rise in temperature brought about different results for the Q10 of Rs but did not cause changes in Rs under different experimental treatments in agroecological systems of the Lhasa.


Introduction
Global surface temperature shows a slower increase during the day than at night [1]; due to the complexity of geographic elements and the interaction of climatic factors, there is still a significant asymmetry in global warming [2]. The magnitude of nighttime temperature increase is significantly higher than daytime [3], and both temperatures' impacts on terrestrial ecological systems are most likely different from each other [4,5]. For instance, a daytime rise in temperature has a stronger effect on net nitrogen mineralization than a nighttime rise does at the global scale [6]. The Qinghai-Xizang Plateau is experiencing noticeable diurnal unsymmetrical rise in temperature [7]. Although increasing numbers of warming studies have discovered the influence of rise in temperature on cold and high-altitude ecological systems on the Qinghai-Xizang Plateau [8][9][10], only a few have explored the influence of diurnal unsymmetrical rise in temperature on cold and high-altitude meadows [11,12]. However, besides these meadows, there are several other cold and high-altitude ecological systems (e.g., agroecological system) [13][14][15][16][17][18][19][20] which have not been researched. Agroecological systems, mainly distributed in semiarid valley districts, are experiencing noteworthy rises in temperature [21][22][23] and are extremely fragile because of high elevation (generally 3000-4500 m), low air temperature and arid/semiarid climatic circumstances in the Qinghai-Xizang Plateau [24][25][26]. Therefore, it is necessary to explore how diurnal unsymmetrical warming can affect the agroecological system.
Soil respiration is closely related to the global carbon cycle [27]. This nonuniformity of daynight temperature increase profoundly affects the response and feedback of terrestrial ecosystems' carbon cycles to global climate change. R s is estimated to be more susceptible to rise in temperature in agroecological systems than other ecological systems (e.g., grassland ecological systems) considering that rural soils have better soil moisture (SM) and ventilation [28]. Several studies have explored the impact of experimental warming on R s in a variety of agroecological systems [29,30]; but there are inconsistent results, reporting both positive [11] and negligible [31,32] effects, on the impact of warming on R s in the Qinghai-Xizang Plateau. Although increasing numbers of studies concentrate on the impact of experimental warming on R s in forest and grassland ecological systems [8,32], only two studies have explored the influence of rise in temperature on R s in agroecological systems of the Qinghai-Xizang Plateau [29,30]. These two studies only explored the influence of rise in temperature on R s during a single growing season, whereas the impact of rise in temperature on R s may change within the year in cold and high-altitude districts [33]. Furthermore, [30] reports the impact of DW+NW treatment on R s only during the daytime (8:00-20:00), while [29] shows the impact of DW +NW treatment on R s during only the daytime (8:00-20:00) and only nighttime (20:00-8:00). Therefore, how diurnal unsymmetrical warming can modify R s in agroecological systems of the Qinghai-Xizang Plateau remains unclear.
To the best of our knowledge, studies of the effect of diurnal unsymmetrical warming on R s in the Qinghai-Tibet Plateau have only been carried out in alpine meadows [11]. However, the response of R s to asymmetric warming in agroecological systems has not been studied. Since the naked barley ecological system is one of the dominant agroecological systems in the Lhasa [34], a diurnal unsymmetrical warming experiment (C: control; DW: 8:00-20:00; NW: 20:00-8:00; DW+NW: 8:00-8:00) was carried out during two growing seasons in 2015-2016 to compare the impacts of the DW, NW and DW+NW treatments on R s and Q 10 of R s in such an ecological system of the Lhasa region.

Research area and experimental design
The study area is located at the Lhasa Agroecosystem Research Station, Tibet Autonomous Region in China(91˚21 0 E, 29˚41 0 N, 3688 m). The annual average temperature is 7.9˚C, and the average annual precipitation is 425 mm. Since the 1970s, the soil in this area has been used for cultivating crops. A number of studies described the research area conditions (e.g., climate traits) [29,30]. Infrared heaters (Kalglo Electronics Inc., Bethlehem, PA, USA) were positioned on the ground approximately 1.7 m in the center of each 2 m×2 m plot and used to enhance soil temperature (T s ) during the two growing seasons of a naked barley ecological system (April 23-August 24 in 2015; April 15-August 16 in 2016). The dimensions and installation heights of infrared heaters, sowing rates and rows of naked barley, and sizes of and distances between sample plots matched those given in a previous study [29]. The field experiment was completely randomized design, with four treatments(C, DW, NW and DW+NW) and four replications per treatment. Both the Lhasa Agroecosystem Research Station and the village leaders gave us permission to conduct our study on their land. All the plants we used were planted by the experimental staff on the station.

Biomass, microclimate and R s measurements
We sampled aboveground and belowground portions of naked barley on August 16, 2016, and rinsed any residual soil off with water. Then, the plant sample was dried at a constant temperature of 65˚C for 48 h; the weight of the aboveground vegetation was the aboveground biomass, and the weight of the root was the underground biomass. We measured soil moisture (SM) and temperature (T s ) (5 cm) during the whole research period in 2015-2016. In the center of each plot, we set T s and SM sensors at a depth of 0.05 m. For each plot, the data of the two sensors was exported via the data logger (HOBO weather station, Onset Computer, Bourne, MA, USA). Data were measured every minute and processed to provide an average value every 5 minutes.
We used LI-8100 (LI-COR Biosciences, Lincoln, NE, USA) to measure daily cycles of R s during the entire research period in 2015-2016. The R s observation interval was 2 h during the daytime (8:00 to 20:00) and 3 h and during the nighttime (20:00 to 8:00). A previous study [29] has shown the sample setting of the R s measuring equipment.

Statistical analyses
We evaluated the impacts of warming and the observation date on R s using repeated-measures ANOVA. We used an independent sample t-test to compare the seasonal mean soil respiration for the two growing seasons under the four treatments. We explored exponential correlations between R s and T s using the Eq 1 and Q 10 of R s using the Eq 2 according to previous studies [29,31].
We performed natural-logarithm transformations between R s and SM and then analyzed the correlations between R s , T s and SM using a multiple stepwise regression analysis. We calculated R s and Q 10 for the full day (8:00-8:00), day (8:00-20:00), and night (20:00-8:00), and compared R s and Q 10 of three periods among the 'C','DW', 'NW' and 'DW+NW' treatments with Student-Newman-Keuls multiple comparisons.
All statistical analyses were performed in the SPSS software (version 22.0; SPSS Inc., Chicago, IL, USA).

Impacts of DW, NW and DW+NW treatments on T s , SM and the correlations between R s and T s
The comparisons of T s , SM and the correlations between R s and T s among the C, DW, NW and DW+NW treatments are illustrated in  (Fig 2). The p-values of all equations are significant (p<0.001); the exponential equation provides a good fit for the relationship between temperature and soil respiration.

Impacts of DW, NW and DW+NW treatments on R s , temperature sensibility and plant biomass
The main impact of the observation date and its interaction impact with warming on R s of the all-day treatment were remarkable. While the main impact of warming on R s was insignificant, there was a significant difference in soil respiration intensity between different stages of the growing seasons in 2015 and 2016 (Table 1). Through Student-Newman-Keuls multiple comparison, we found R s of all-day treatment among the C, DW, NW and DW+NW treatments had no significant difference. The comparisons of R s among the C, DW, NW and DW+NW treatments are illustrated in Fig 3, while the comparison of the seasonal mean soil respiration for the two growing seasons under the four treatments is illustrated in Fig 4. Through an independent sample t-test, it was found that there was no significant difference in seasonal mean soil respiration between the two years.
The effects of the four treatments (C, DW, NW and DW+NW) on daytime and nighttime soil respiration showed no significant difference (Fig 5A). The effects of four treatments (C, DW, NW and DW+NW) on the daytime and nighttime Q 10 were different, while those on the daytime Q 10 had no significantly difference; Compared with the C treatment, DW, NW and DW+NW had significant inhibitory effects on the nighttime Q 10 , with all treatments having no significant difference between them. From the all-day Q 10 , only C and DW showed a significant difference (Fig 5B). From Fig 2,   difference between the C and DW treatments was the largest, consistent with the multiple comparison results of all-day Q 10 between the four treatments ( Fig 5B). The DW, NW and DW+NW treatments had negligible impacts on plant biomass (Fig 6).  and T s exceeded those between R s and SM of the DW, NW and DW+NW treatments. Compared with SM, T s can better explain the change in R s . The Q 10 value decreased significantly with increasing T s across all the experimental plots, while the Q 10 value only tended to increase with increasing SM (Fig 7).

Discussion
We have discovered that the degree of soil drought produced by the DW+NW treatment was substantially larger than that by the DW and NW treatments, which possibly originated from the fact that the DW+NW treatment brought about a larger increase in T s than did the DW and NW treatments [29,31,32,35]. Our findings suggest that diurnal unsymmetrical warming will not always trigger unsymmetrical results on plant biomass in cold and high-altitude districts, which is possibly due to unsuitable increase in T s , warming-induced decline in SM and an increase in plant respiration [32,[35][36][37][38][39][40]. Our findings suggest that warming can modify the relative impacts of T s and SM on R s in cold and high-altitude districts [29,31,41]. Vegetation growth and soil microorganism activity are strongly influenced by low temperatures [8,10,42,43], thus rise in temperature is estimated to amplify R s in cold and high-altitude districts [8]. However, several previous studies noted that rise in temperature had no effect on R s in agroecological systems and cold and high-altitude meadows in the Qinghai-Xizang Plateau [29,32,44], a result which was additionally supported by this study. All of these findings entailed that R s most likely has no significant response to warming in cold and high-altitude districts. Our findings did not follow the lines of earlier meta-analyses which showed that R s increased during the warming treatments [45,46].
We have discovered that the Q 10 of R s between the DW and NW treatments were diverse, and moreover, that DW and NW treatments can lead to differing effects in the daytime and nighttime Q 10 . This indicated that Q 10 was not only related to the different ecosystems and the magnitude of temperature increase, but different warming methods as well. The all-day Q 10 values were 1.69, 1.55, 1.66 and 1.58 for the C, DW, NW and DW+NW treatments, respectively. Compared with C treatment, DW and NW Q 10 values decreased by 9% and2%, respectively, but the DW + NW Q 10 decreased by 7% indicating that the impact of DW and NW was not a simple additive relationship. As can be understood from  all-day Q 10 values, while NW and DW+NW had no significant change; this further proves that these two processes were not simple additive relationships. We even speculated that DW and NW may have antagonistic effects; as a consequence, the asymmetry of daytime and nighttime rises in temperature will possibly bring about unsymmetrical results for the temperature sensibility of R s in cold and high-altitude rural soils.
Q 10 can be used to reflect the relationship with temperature, as its value is closely related to temperature. From Fig 7, it can be seen that Q 10 is negatively correlated with temperature, which is consistent with previous studies [47][48][49]. Kirschbaum[50] analyzed the results of different studies and found that Q 10 was very high at low temperatures and relatively stable at high temperatures. Luo [51]found that the Q 10 value of the treatment without warming was significantly higher than the Q 10 value of the experimental warming treatment in a study on a cold and high-altitude grassland in North America. Moreover, we found that the Q 10 value of soil respiration during daytime treatment was lower than that of nighttime (Fig 5B), which further confirmed the negative correlation between temperature and Q 10 .
The minor differences of R s among the C, DW, NW and DW+NW treatments are possibly due to the following reasons. First, there were minor differences of BGB among the C, DW, NW and DW+NW treatments, and R s was positively correlated with BGB [52]. Next, although the DW+NW treatment brought about a larger rise in T s , it triggered more soil drought than did the DW and NW treatments which in turn was able to lessen the impact of rise in temperature on R s . Third, the warming degrees of the DW, NW and DW+NW treatments possibly did not reach the optimum temperature which could lead to maximum increase in R s [32]. Lastly, only the DW treatments, not the NW and DW+NW treatments, dramatically lowered the Q 10 of R s , which in turn was able to weaken the impact of rise in temperature on R s [53].

Conclusions
This study explored how diurnal unsymmetrical warming was able to modify R s and temperature sensibility of R s in an agroecological system within the Qinghai-Xizang Plateau during two growing seasons in 2015-2016. The R s differences among the C, DW, NW and DW+NW treatments were able to be neglected. The effects of DW, NW and DW+NW treatments on Q 10 of R s during daytime and nighttime were different; they both had no effects on daytime soil respiration sensitivity, but significant inhibitory effects on nighttime soil respiration sensitivity. Moreover, we speculate that DW and NW have antagonistic effects on Q 10 of R s .
Through warming experiments, we found that the soil respiration under the diurnal unsymmetrical warming has different sensitivities while the magnitude of warming caused by global warming differed between day and night. The experiment explores the effects of the asymmetric warming mechanism on soil respiration, providing a theoretical basis for better simulation and prediction of soil respiration so as to better explain the response and feedback of terrestrial ecosystems' carbon cycle to global climate change.