| Terrestrial ecosystem carbon exchange process and its carbon balance and evaluation is one of the important contents of the carbon flux research in China. Through research the flux observation in wetland ecological system under the background of Ebinur desert in Xinjiang province, based on dealing with flux data and comparative analyzing the results with the mainstream processing software, this study discussed the data interpolation model and characteristics in the study area, the change rules and energy balance characteristics of NEE desert ecological system are also analyzed in time scale, revealing the spatio temporal variations characteristics of carbon flux contribution area in Ebinur Lake desert ecosystem, delve into the influence and contribution of plants, soil and waters on the underlying surface to the carbon budget, and at last, to estimate the gross primary productivity(GPP)of Ebinur Lake desert ecosystem, and the numerical simulation was performed through the site scale model(CEVSA2). This research would make up the lack of observation data on conducting research on “carbon source/sink” in arid region, and providing the theoretical basis and foundation for accurate estimation of carbon source/sink function change and carbon cycle process mechanism research of typical ecosystems in arid region.The main results and conclusions were summarized as follows:(1)The carbon flux of Ebinur Lake desert ecosystem strongly affected by environment, more outliers of the original CO2 concentration and carbon flux concentrated in winter(proportion was more than 40 %), and carbon flux under different weather condition(sunny, cloudy, rainy, windy, snowy and frost)all showtheirselves regularity. After the parameters revised, Edi Re and Edd Pro software could effectively process the flux data, and the results were similar(R2=9776,P<0.05), and there were only exist differences in negative territory, and the degree of fitting is relative less(R2=0.7670,P<0.05); There werelittle difference in sensible heat flux calculated by those two software, and latent heat flux calculated by Eddy Pro was approximately twice times than Edi Re. Calculated data by Eddy Pro was more reasonable and less volatile, and data quality grade was relatively weak than Edi Re. which can improve the continuity and integrity of data, therefore, Eddy Pro software is the optimal choice for flux data processing in Ebinur Lake desert ecological system.(2)Valid data of the whole year carbon flux accounted for 54.42 % in 2012, the proportion for empty data, sensors abnormal, check out and unreasonable threshold of the totaldata were 18.21 %, 9.43 %, 2.72 % and 15.22 %, respectively; April 23 rd and October 18 th were the critical days for plus or minus shift of ecosystem net CO2 exchange capacity( NEE); using the Michaelis-Menten response model to interpolatethe diurnal NEE during growing season, the maximum initial light energy utilization ratio(α), saturated light intensity ecosystem carbon assimilation capacity( Pmax), diurnalecosystem respiration rate( Reco,d) were 0.01511 mg CO2?μmol-1?quantum-1、38.640 and 2.599 μmol?m-2?s-1,and they were exist July, June and July, respectively. The maximumα, Pmax and Reco,d were 0.01825 mg CO2?μmol-1?quantum-1, 20.540 and 2.2180 μmol?m-2?s-1 from July to Sept. in 2014; NEE and temperature(Ta)at night in growing season meet the Lloyd & Taylor respiration equation, andthe index fitting was more ideal(R2=0.6222,P<0.05), respiratory rate for ecological system at 25 ℃ was 3.438 μmol?m-2?s-1;Using the van ’t Hoff index equation to fit and interpolate the data in non-growing season, the change range of NEE was only from 0 to 2.5 μmol?m-2?s-1 under the soil temperature at-15~22 ℃.(3)Daily averages of turbulence carbon flux(Fc)in Growing season was-2.519 μmol?m-2?s-1, diurnal Fc of non-growing season swing, and its average value was 1.803 μmol?m-2?s-1, which showed that the space from the above the plant canopy to the interface under the atmosphere, which absorb CO2 in growing season and release it in non-growing season. Fc accounted for NEE was more than 95 %, and CO2 Store flux fluctuate between-1.0 and +1.0 μmol?m-2?s-1; The maximum monthly average of NEE in non-growing season appeared in April, and the value was 1.019 μmol?m-2?s-1, NEE from May to Sept. were all positive, and the maximummonthly average appear in June, its value was 2.025 μmol?m-2?s-1, and the minimum monthly average NEE in Sept., the value was-0.638 μmol?m-2?s-1,monthly average NEE from July to Sept. in 2012 are all less than 2014, and there were great difference in August, and the latter was 1.89 times than the former, which indicated that CO2 absorption capacity of Ebinur Lake desert ecosystem in 2012 than weaker in 2014,there was the biggest ecosystem carbon accumulation in August, 2014.(4)The study area is prevailing southeast and northwest wind in 2012 and 2014; On the whole, the atmospheric stability presentneutral characteristics, whose maximum value was 0.842, and appear in September, 2014, the biggest flux contribution area in the direction of 270o~360o in 2012, the flux contribution area in the direction of 70o~360o are all had a slight increase in 2014, the four typical daily energy balance components are characterized by high value in day-time and low value in night-time, high value at noon and low value at early hours in the morning, which show unimodal carve, and sensible heat flux as the major energy-consuming way in the spring, while latent heat is greater than the sensible heat in summer, autumn and winter; On the whole, at the yearly scale, latent heat flux showed increase in spring and summer, while reduce in autumn and winter, and latent heat flux is swing in summer, and moisture is key factor of theratio of sensible heat and latent heat flux ratio; daily change characteristics of residual energy in desert ecosystem has a significantly certain period of time; Compared to different underlying surfaces in arid areas, latent heat flux of desert ecosystem has higher ratio in energy expenditures, as well as the Bowen ratio is less than the average level in arid areas.(5)The dominant plants of underlying surface are Halostachys caspica, Phragmites australis and Haloxylon ammodendron, order for net photosynthesis rate( Pn) of those three plantswere: Halostachys caspica>Phragmites australis>Haloxylon ammodendron, namely, arbor>shrub>herb, Pn of arbor significantly higher than shrub and herb in July, while there were no significant difference among those three plants; Change law of plant Pn and NEE are similar basically, the order for contribution of plant respiration to NEE was: Haloxylon ammodendron>Phragmites australis>Halostachys caspica; The contribution of dominant plants Pn to NEE was 42.31 % in day(7:00~17:00), while the contribution was 54.66 % at night(19:00~5:00 in the morning the next day); And the contribution was 46.3 % at night in Sept..(6)Air temperature, air relative humidity and potential evaporation change trend were consistent in forest land and waters of underlying surface, and there are little difference between them, while there exist great difference in wind speed between them; turbulence carbon flux, vapor flux and NEE above the waters were greater than forest land, The former were 2.08, 1.97, and 2.30 times than the latter, respectively; Soil volumetric water content decreases with distance from the banks of the river, and soil temperature change slightly, while soil electrical conductivity at 60 cm depth change great; On the whole, the soil respiration rate increases with distance from the banks of the river, the contribution of soil respiration to NEE in ecosystem at night was 82.9 % in July, while dark respiration of plants was 17.1 %, and the contribution of soil respiration to NEE in ecosystem at night was 26.93 %, while dark respiration of plants was up to 73.07 %, in Sept., indicated that ecosystem respiration at night is given priority to plant dark respiration in Sept..(7)Overall, the gross primary productivity(GPP)of the Ebinur lake desert ecosystem showed a trend from increase to decreases during growing season in 2012, GPP was 1.109 g C?m-2?d-1 in the late April, maximum GPP appeared in the early and middle of July, and GPP near to 0 in the early and middle of October, while on the whole, the change trend of ecosystem cumulative daily NEE is in contrast to the GPP; ecosystem NEE, Reco and GPP were-229.34, 364.89 and 594.23 g C?m-2 during growing season in 2012, respectively; GPP during July to Sept. were 349.99 and 365.81 g C?m-2 in 2012 and 2014, respectively, the higher value of NEE in 2014 than in 2012, the difference was 15.82 g C?m-2; GPP simulation data for “Carbon Exchange in the Vegetation-Soil-Atmosphere”(CEVSA2)model was greater than measured data during non-growing season in 2012, while which was less during growing season, GPP change range was variable in 2014, the correlation coefficient of measured data and simulation data for GPP were 0.6177 and 0.7082 in 2012 and 2014, respectively, CEVSA2 model could well simulate the desert ecosystem GPP in Ebinur Lake basin. |
PDF Full Text Request