The Research On Carbon Budget In Forest Ecosystems Based On Wireless Sensor Networks

As a significant observational and measuring approach in earth system science,wireless sensor networks(WSN)has been widely used in environmental monitoring.To accurately simulate and predict how terrestrial ecosystem functions respond to climate changes,it is necessary to better understand the structure and process of carbon cycle.WSN can continuously monitor environmental factors in a fine-scale region,which is helpful to estimate the carbon flux and explain compositions and distribution of carbon budget in a forest at a site scale.Although WSN has the advatanges in carbon cycling studies,it just has been used in this field.Therefore,there still existed some problems should be stressed,such as the evaluation of WSN statistics,their correction method and the calculation method for the carbon flux.Bsed on the high-precision meteorological gradient system and CO2/H2O profiler system into observing and measuring statistics,this study.evaluates the accuracy of observed results(including temperature,humidity and CO2 concentration)of the wireless sensors,analyzes the source of error and constructs a corresponding correction model.From this point of theories,statistic requirements and simulation verification,this study also test the applicability of different calculation methods for carbon flux in the process of estimating the carbon flux of forest ecosystem via the statistics collected by wireless sensors.Bsed on the vertical and horizontal observations via wireless sensors at Tianmushan and Anji Experimental Area,this study uses corrected wireless sensor data and suitable carbon flux estimation model to study the structure and constitution of carbon flux in forest ecosystems as well as the differences among various types of forests.Some conclusions are drawn as follows:(1)The total accuracy of the statistics collected by temperature wireless sensors is 87%(R2=0.94)with the measured values higher at day and lower during night than the reality.Moreover,the bias tend to be larger on sunny days than those on cloudy days,accounting for 51%of the total bias.The bias mainly induced from the plastic shell of the sensors.The solar radiation certainly increases the temperature of the shell,which exerts further influence on the observed results.The total accuracy of the data collected by humidity wireless sensors is 82%(R2=0.89).The higher measurements are found in the observed values than the reality.The measurement showed higher bias in nighttime than those in daytime,and higher bias in sunny days than those in cloudy days.The bias in the nightime of sunny days contributed 64%of the total bias,mainly because in such conditions,dew is more likely to form on the sensor shells,thus making their inner parts moister than the surrounding air.The total accuracy of the data collected by CO2 concentration wireless sensors is 95.1%(R2=0.74).Due to the operating principle of the sensors,relative air humidity is the major cause for bias.Moister environment will lead to higher observational values.Taking humidity as a variable,an error correction model is constructed and thus the measuring accuracy of temperature,humidity and CO2 concentration is respectively increased to 91%,88%and 96.4%.(2)According to the eddy covariance method,when the sampling frequency of CO2 concentration and the corresponding wind speed decrease,there will be decline in the diurnal variation of carbon flux time series and growth in the random fluctuation and the estimation error of the flux.Generally speaking,the carbon flux derived from wireless sensor CO2 concentration data based on the eddy covariance method is reliable when the sampling frequency of CO2 concentration and vertical wind velocity is the highest one(0.17 hz)for the wireless sensor and the spatial positions of the sensors are close together.When WSN is used in observing and measuring the vertical gradient of environmental factors,the diurnal variation of temperature,humidity,CO2 concentration and wind speed at different vertical heights is mainly affected by the metabolic activity of the underlying surface vegetation,the canopy height and the solar radiation.Based on aerodynamics,an estimation of carbon flux is achieved using the vertical gradient statistics of environmental factors observed by wireless sensors.It shows that,though carbon flux is basically consistent with eddy flux at the height of 17-30m above the vegetation canopy,the estimation result is a little higher than the reality,paricularly during nighttime,which may result from the CO2 accumulation caused by insufficient air flow.The relationships between environmental factors and forest carbon budget components show that forest GPP is well correlated with daytime anomalies of canopy CO2 concentration and temperature,and leaf and soil respiration rates show a strong correlation with temperature.When using the data of environmental factors collected by wireless sensors to calculate carbon flux based on compartment model of forest carbon cycle,its diurnal variation can be clearly demonstrated,but it is hard to capture the specific changes and inevitable to get an overestimate at daytime.(3)In summer and witner,the surface layer is performed as a carbon source,the canopy layer is performed as a carbon sink in the daytime and as a carbon source in the nighttime for the vertical pattern of carbon flux in the broadleaf forest and bamboo forest.The effect of canopy-above layer is determined by the carbon budget of land surface vegetation.In Tianmushan broadleaf forest,there is scrub by vegetation on the ground,which has a strong photosynthetic capacity in autumn and winter because of its adaptability to the phenology of canopy vegetation.The carbon uptake of the scrub by vegetation compensates apart of soil respiration CO2 emission,that make a diurnal variation in surface-layer NEP(net ecosystem productivity)as an inverted U-shaped.For the horizontal pattern of carbon flux in the broad leaf forest and bambooforest,the spatial differences of environmental factors in summer are higher than those in winter.Additionally,the spatial differences of environmental factors in the daytime are greater than those in the nighttime,and the spatial heterogeneities show a better synchronicity between carbon flux and temperature in a diurnal cycle.Tianmushanbroadleaf forest and Anjibamboo forest are all carbon sink in both summer and winter,and the carbon sink intensity and sequestration effectiveness in summer are greater than those in winter.The carbon sequestration effectiveness of the broad leafforest is higher than that of the bamboo forest.The ecosystem respiration in winter is weaker than that in summer,but accounts for a higher proportionof GPP.In winter,the ratio of leaf respiration to GPP in the bamboo forest is greater than that in the broadleaf forest,indicating that it consumes large amounts of carbon to maintain the relatively high photosynthetic capacity for the bamboo leaves.According to differences of carbon flux compositions between the two forestecosystems in this study,it indicates that the sub-tropical broadleaf forest ecosystem has a kind of relatively conservative ecological strategies,while the bamboo forest has a kind of aggressive ecological strategies,which is a physiological characteristic for an invasive species with strong growing and expansion ability.