| In the context of global change,the carbon,water and energy cycle of the Earth system is attracting more and more attention.Several methods have been used for measuring carbon fluxes at the ecosystem scale.The eddy covariance(EC)method is a representative of the micro-meteorology method,which directly measures the net CO2 exchange between terrestrial ecosystems and the atmosphere and simultaneously measures the fluxes of water and energy.As the most complex types of terrestrial ecosystems,the forest ecosystem has the most rich tree species,highest canopy and largest spatial heterogeneity,which make large difficulties in the determination of carbon,water and energy budgets.Because the application of EC method in forest carbon,water and energy fluxes belongs to the interdisciplinary of ecology,forestry and micrometeorology,the research of its theoretical and technical problems lags behind the needs of practical application.In order to fulfill the role of EC-flux networks in the global carbon and water cycle and climate change,it is extremely urgent to solve the key problems in its theoretical and practical applications.The Maoershan flux tower was about 400 m above sea level and was located on the NE-slope of the small valley.So it is not only representative of the climate,topography and forest type in the mountainous area of the northeastern of China,but also has a high similarity with the topography of the Daxinganling forest.It also has great enlightenment for the global mountain forest.In this thesis,using the specific dataset during 2008 to 2017,the temperate deciduous broadleaved forest that located on a valley-slope at the Maoershan Station was taken as an example to study four relatively-independent but closely-related key issues of EC flux measurements:estimation CO2 storage flux(Fs)using a single-profile,coordinate rotation(ultrasonic anemometer tilt correction),surface heating effect and its correction of the open-path analyzer,radiation measurement method and its impact on radiation and energy balance over sloping topography.I got the following results:(1)The errors in Fs relevant to the vertical distribution of sampling levels and window sizes of averaging time of CO2 dry molar fraction and their effects on NEE were assessed using the standardized major axis method.The CO2 storage per unit height typically decreased with the height increasing,suggesting that the below-canopy layer need a higher spatial resolution.CO2 storage could be underestimated by up to one third based only on the tower-top measurement.The uncertainty(standard deviation)of the Fs decreased with the length of the averaging time window increasing.However,taking time averaging of CO2 dry molar fraction(?c)caused significant underestimate of Fs,and consequently led to significant underestimates of CO2 uptake and release at a 30-min time scale.During the growing season,the underestimation of carbon uptake by the 30-min averaging of ?c was up to 5.4?mol m-2 s-1(6:30),and in the evening,the underestimation of CO2 release was estimated to be about 3.9?mol m-2 s-1(19:30).The results highlight that appropriately combiming spatial resolution and temporal resolution(response time for a whole sampling of all levels)is essential to improving the Fs estimates with the existing CO2 profile systems in forest ecosystems.(2)The effects of various tilt-correction methods on sensible heat flux(H),latent heat flux(LE),CO2 eddy fluxes(Fc),and energy balance,friction velocity were comprehensively examined using the standardized major axis method.Compared with the unrotated fluxes,the double rotation(DR),planar fit(PF),vertical velocity unbiased planar fit(NBPF),and monthly planar fit(MPF)increased the H by 1.5%,3.8%,3.1%,and 3.3%,respectively,while the triple rotation(TR)and monthly sector-wise plane fit(MSWPF)decreased the H by 1.3%and 7.1%,correspondingly.The NBPF and DR increased the LE by 1.0%and 0.4%,respectively;the PF also increased it by 0.671 W m-2;the MPF decreased it by 0.4%,but increased the intercept by 0.747 W m-2;in contrast,the MSWPF and TR decreased the LE by 4.5%and 15.1%,respectively.The coordinate rotation reduced the Fc by 6.0%(DR)-12.1%(TR).Among the studied methods,the PF increased the energy balance closure by about 2%,whereas the TR decreased it by about 6%.In the light of friction velocity and vertical velocity,the TR and NBPF were inappropriate for anemometer tilt correction.Considering the accuracy and practical application of EC measurements,I recommended the PF as the first choice for the Maoershan flux tower site,followed by the DR.These results provide a reference for anemometer tilt correction for EC flux observations under complex terrain conditions.(3)The surface(bottom lens,top lens and strut)temperature of LI-7500 was directly measured using T-type thermocouples with a bare-wire diameter of 0.127 mm.Simultaneously,high frequency air temperature in the optical path of LI-7500 and the ambient air using K-type fine-wire thermocouples with a bare-wire diameter of 0.079 mm,and explored the assessment and correction methods of the LI-7500 surface heating effect.The surface-heating of LI-7500 varied with the day and night alteration,season and location:higher in the daytime than in the nighttime,larger in winter than in summer,and the bottom lens>spar>top lens,with the winter bottom-window heating higher than 5?.The surface heating estimated by the seven methods resulted in a large difference in the pattern of the increase of H and the consequent increase of the CO2 turbulence flux.The Fc-errors due to surface heating could reach 1.0 ?mol m-2s-1 in summer and typically about 2.0 ?mol m-2 s-1 in winter.Using the combination of the measured surface temperature with the T-type thermocouple and the Nobel equation(TS)method as the reference,both Burba univariate and multivariate linear equations trended to overestimate the heating effect;The site-specific univariate linear model and the thermocouple-based increase of H were similar to the TS method in summer;The site-specific univariate linear model was closed to the TS method in winter;The method of using the fine-wire thermocouples for determining the sensible heat flux had a high sensitivity and a high nois;.However,the simulation model of the sensible heat flux of the fine-wire thermocouples could not reflect the impact of microclimatological factors very well.Aerodynamic-based nonlinear regression models might improve the simulation accuracy.(4)Using a temperate deciduous forest on a moderate inclined(9°)northwest-facing slope as a case study,I quantified the slope effect on net radiation(Rn)and its components as well as the albedos in shortwave(Rs),photosynthetically active(PAR),and near-infrared(NIR)radiations from early May to early October 2016.The influence of radiometer-orientation on the measured energy balance closure by an eddy covariance system was also assessed.The effect of sensor-orientation on the diurnal variations in radiations was larger for incoming than outgoing radiations,and larger in Rs than longwave radiation(L).Compared with the slope-parallel radiometer,the horizontal sensor overestimated the incident Rs by 7%,the incoming NIR by 10%,and the incoming PAR by 2%;while the horizontal sensors underestimated the reflected Rs by 4%and NIR by 5%.Summing these components,horizontal sensor overestimated the Rn by 9%.Converting the horizontally-measured incident radiation to slope-surface reduced a quarter of the biases on incoming Rs and Rn.The horizontal radiometers generally underestimated the daily means of albedos,particularly in the NIR waveband.Measuring the Rn with slope-parallel radiometer improved the energy balance ratio(EBR)by 8.3%,but correcting the slope-effect on horizontally-measured incident Rs only improved it by 1.4%.A mini-review of 10 sites globally indicated that,the horizontal radiometer underestimated the morning radiation and overestimated the afternoon radiation on east-facing slope,and opposite on the west-facing slope.The horizontal sensor underestimated(overestimated)the EBR on north-facing(south-facing)slopes,with an sensitivity of EBR due to horizontal-radiometers as high as 1.1%per degree of inclination angle.The estimations of light response parameter of NEE was also affected by the way the radiameter is installed.Although this paper was only a case study at the Maoershan site,its topography,vegetation and climate had a wide representativeness in flux sites in Northeast China.And some general rules were obtained from comparisons with the results of other sites.These novel conclusions effectively explored and supplemented the EC theory and technology,and would promote the development of this cross-cutting field.There is a need to further evaluate the cumulative error on the annual scale in the future.Finally,the assessment of EC-errors also needs the cross-validation of the results of other independent measurements taken on the ground.The detailed biometric and chamber-based CO2 budget methods can be used to improve the accuracy of EC fluxes. |
PDF Full Text Request