Tower Footprint Research Articles

Sensitivity of surface flux simulations to hydrologic parameters based on an uncertainty quantification framework applied to the Community Land Model

Uncertainties in hydrologic parameters could have significant impacts on the simulated water and energy fluxes and land surface states, which will in turn affect atmospheric processes and the carbon cycle. Quantifying such uncertainties is an important step toward better understanding and quantification of uncertainty of integrated earth system models. In this paper, we introduce an uncertainty quantification (UQ) framework to analyze sensitivity of simulated surface fluxes to selected hydrologic parameters in the Community Land Model (CLM4) through forward modeling. Thirteen flux tower footprints spanning a wide range of climate and site conditions were selected to perform sensitivity analyses by perturbing the parameters identified. In the UQ framework, prior information about the parameters was used to quantify the input uncertainty using the Minimum‐Relative‐Entropy approach. The quasi‐Monte Carlo approach was applied to generate samples of parameters on the basis of the prior pdfs. Simulations corresponding to sampled parameter sets were used to generate response curves and response surfaces and statistical tests were used to rank the significance of the parameters for output responses including latent (LH) and sensible heat (SH) fluxes. Overall, the CLM4 simulated LH and SH show the largest sensitivity to subsurface runoff generation parameters. However, study sites with deep root vegetation are also affected by surface runoff parameters, while sites with shallow root zones are also sensitive to the vadose zone soil water parameters. Generally, sites with finer soil texture and shallower rooting systems tend to have larger sensitivity of outputs to the parameters. Our results suggest the necessity of and possible ways for parameter inversion/calibration using available measurements of latent/sensible heat fluxes to obtain the optimal parameter set for CLM4. This study also provided guidance on reduction of parameter set dimensionality and parameter calibration framework design for CLM4 and other land surface models under different hydrologic and climatic regimes.

Retrieval of vertical LAI profiles over tropical rain forests using waveform lidar at La Selva, Costa Rica

This study explores the potential of waveform lidar in mapping the vertical and spatial distributions of leaf area index (LAI) over the tropical rain forest of La Selva Biological Station in Costa Rica. Vertical profiles of LAI were derived at 0.3m height intervals from the Laser Vegetation Imaging Sensor (LVIS) data using the Geometric Optical and Radiative Transfer (GORT) model. Cumulative LAI profiles obtained from LVIS were validated with data from 55 ground to canopy vertical transects using a modular field tower to destructively sample all vegetation. Our results showed moderate agreement between lidar and field derived LAI (r2=0.42, RMSE=1.91, bias=−0.32), which further improved when differences between lidar and tower footprint scales (r2=0.50, RMSE=1.79, bias=0.27) and distance of field tower from lidar footprint center (r2=0.63, RMSE=1.36, bias=0.0) were accounted for. Next, we mapped the spatial distribution of total LAI across the landscape and analyzed LAI variations over different land cover types. Mean values of total LAI were 1.74, 5.20, 5.41 and 5.62 over open pasture, secondary forests, regeneration forests after selective-logging and old-growth forests respectively. Lastly, we evaluated the sensitivities of our LAI retrieval model to variations in canopy/ground reflectance ratio and to waveform noise such as induced by topographic slopes. We found for both, that the effects were not significant for moderate LAI values (about 4). However model derivations of LAI might be inaccurate in areas of high-slope and high LAI (about 8) if ground return energies are low. This research suggests that large footprint waveform lidar can provide accurate vertical LAI profile estimates that do not saturate even at the high LAI levels in tropical rain forests and may be a useful tool for understanding the light transmittance within these canopies.

Seasonal variation of CH4 emissions from central California

We estimate seasonal variations in methane (CH4) emissions from central California from December 2007 through November 2008 by comparing CH4 mixing ratios measured at a tall tower with transport model predictions based on a global 1° a priori CH4emissions map (EDGAR32) and a 10 km seasonally varying California‐specific map, calibrated to statewide by CH4emission totals. Atmospheric particle trajectories and surface footprints are computed using the Weather Research and Forecasting and Stochastic Time‐Inverted Lagrangian Transport models. Uncertainties due to wind velocity and boundary layer mixing depth are evaluated using measurements from radar wind profilers. CH4signals calculated using the EDGAR32 emission model are larger than those based on the California‐specific model and in better agreement with measurements. However, Bayesian inverse analyses using the California‐specific and EDGAR32 maps yield comparable annually averaged posterior CH4emissions totaling 1.55 ± 0.24 times and 1.84 ± 0.27 times larger than the California‐specific prior emissions, respectively, for a region of central California within approximately 150 km of the tower. If these results are applicable across California, state total CH4 emissions would account for approximately 9% of state total greenhouse gas emissions. Spatial resolution of emissions within the region near the tower reveal seasonality expected from several biogenic sources, but correlations in the posterior errors on emissions from both prior models indicate that the tower footprints do not resolve spatial structure of emissions. This suggests that including additional towers in a measurement network will improve the regional specificity of the posterior estimates.

Large amplitude spatial and temporal gradients in atmospheric boundary layer CO2mole fractions detected with a tower‐based network in the U.S. upper Midwest

This study presents observations of atmospheric boundary layer CO2mole fraction from a nine‐tower regional network deployed during the North American Carbon Program's Mid‐Continent Intensive (MCI) during 2007–2009. The MCI region is largely agricultural, with well‐documented carbon exchange available via agricultural inventories. By combining vegetation maps and tower footprints, we show the fractional influence of corn, soy, grass, and forest biomes varies widely across the MCI. Differences in the magnitude of CO2 flux from each of these biomes lead to large spatial gradients in the monthly averaged CO2mole fraction observed in the MCI. In other words, the monthly averaged gradients are tied to regional patterns in net ecosystem exchange (NEE). The daily scale gradients are more weakly connected to regional NEE, instead being governed by local weather and large‐scale weather patterns. With this network of tower‐based mole fraction measurements, we detect climate‐driven interannual changes in crop growth that are confirmed by satellite and inventory methods. These observations show that regional‐scale CO2 mole fraction networks yield large, coherent signals governed largely by regional sources and sinks of CO2.

Estimating the net ecosystem exchange for the major forests in the northern United States by integrating MODIS and AmeriFlux data

The eddy covariance technique provides long-term continuous monitoring of site-specific net ecosystem exchange of CO2 (NEE) across a large range of forest types. However, these NEE estimates only represent fluxes at the scale of the tower footprint and need to be scaled up to quantify NEE over regions or continents. In the present study, we expanded a method developed previously and generated a new NEE model exclusively based on the Moderate Resolution Imaging Spectroradiometer (MODIS) products, including enhanced vegetation index (EVI), land surface water index (LSWI), land surface temperature (LST) and Terra nighttime LST′. This method, in our previous research, provided substantially good predictions of NEE and well reflected the seasonal dynamics of the deciduous broadleaf forest at the Harvard forest site. Studying NEE of forests in the middle-latitude regions of the Northern Hemisphere is significant because it may help to understand the ‘missing carbon sink’ from terrestrial ecosystems. In this study we selected eight eddy flux sites to represent the major forest ecosystems in the northern United States. Compared with the model based on a single site, we also established the general models that apply to evergreen needleleaf forest (ENF) and deciduous broadleaf forest (DBF), respectively. The results showed that our simpler model based entirely on MODIS products promised well to estimate NEE by the eddy covariance technique. The modeled annual mean NEE from DBF deviated from the measured NEE by 44.4%, whereas the modeled NEE from ENF was extremely close to the measured NEE within 5.5%. In the end, we also validated both general models for ENF and DBF using independent flux sites. It demonstrated this method performed well for estimating NEE.

Characteristics of the flux of isoprene and its oxidation products in an urban area

[1] We measured the mixing ratios and fluxes of isoprene and its oxidation products, methacrolein (MACR) and methyl vinyl ketone (MVK) from a tall flux tower in metropolitan Houston, Texas, during summertime using a relaxed eddy accumulation (REA) system combined with a dual-channel GC-FID instrument. We show that isoprene was affected dominantly by biogenic emission sources during daytime, but also that tail-pipe emission sources (alongside 2-pentenes) are contributing during the rush hours and at night. The observed daytime mixing ratios of isoprene were much lower than over forested areas due to a comparatively low density of isoprene emitting trees in the tower's footprint area. Daytime isoprene fluxes displayed the expected light and temperature driven emission characteristics, and a detailed isoprene emissions model explained average fluxes fairly well. Our investigation of isoprene's oxidation products MACR and MVK showed that both anthropogenic and isoprene oxidation sources exist for MACR, while MVK was strongly dominated by isoprene oxidation between its emission and sampling points due to presumably high local OH radical concentrations. While biogenic emission modeling appears to work well for this urban environment when supplied with accurate input data, emission inventories used for ozone modeling may need to consider tailpipe isoprene (and MACR) emissions to properly account for urban concentrations, particularly higher morning abundances.

Contributions of primary and secondary biogenic VOC tototal OH reactivity during the CABINEX (Community Atmosphere-Biosphere INteractions Experiments)-09 field campaign

Abstract. We present OH reactivity measurements using the comparative reactivity method with a branch enclosure technique for four different tree species (red oak, white pine, beech and red maple) in the UMBS PROPHET tower footprint during the Community Atmosphere Biosphere INteraction EXperiment (CABINEX) field campaign in July of 2009. Proton Transfer Reaction-Mass Spectrometry (PTR-MS) was sequentially used as a detector for OH reactivity and BVOC concentrations including isoprene and monoterpenes (MT) for enclosure air. Therefore, the measurement dataset contains both measured and calculated OH reactivity from well-known BVOC. The results indicate that isoprene and MT, and in one case a sesquiterpene, can account for the measured OH reactivity. Significant discrepancy between measured OH reactivity and calculated OH reactivity from isoprene and MT is found for the red maple enclosure dataset but it can be reconciled by adding reactivity from emission of a sesquiterpene, α-farnesene, detected by GC-MS. This leads us to conclude that no significant unknown BVOC emission contributed to ambient OH reactivity from these trees at least during the study period. However, this conclusion should be followed up by more comprehensive side-by-side intercomparison between measured and calculated OH reactivity and laboratory experiments with controlled temperature and light environments to verify effects of those essential parameters towards unknown/unmeasured reactive BVOC emissions. This conclusion leads us to explore the contribution towards ambient OH reactivity (the dominant OH sink in this ecosystem) oxidation products such as hydroxyacetone, glyoxal, methylglyoxal and C4 and C5-hydroxycarbonyl using recently published isoprene oxidation mechanisms (Mainz Isoprene Mechanism II and Leuven Isoprene Mechanism). Evaluation of conventionally unmeasured first generation oxidation products of isoprene and their possible contribution to ambient missing OH reactivity indicates that the ratio of OH reactivity from unmeasured products over OH reactivity from MVK + MACR is strongly dependent on NO concentrations. The unmeasured oxidation products can contribute ~7.2 % (8.8 % from LIM and 5.6 % by MIM 2 when NO = 100 pptv) of the isoprene contribution towards total ambient OH reactivity. This amount can explain ~8.0 % (9.7 % from LIM and 6.2 % from MIM 2) of missing OH reactivity, reported by Di Carlo et al. (2004) at the same site. Further study on the contribution from further generation of unmeasured oxidation products is needed to constrain tropospheric photochemical reactivity of BVOC that have important implications for both photochemical ozone and secondary organic aerosol formation.

Relating the X-band opacity of a tropical tree canopy to sapflow, rain interception and dew formation

During summer and autumn 2007, a 11 GHz microwave radiometer was deployed in an experimental tree plantation in Sardinilla, Panama. The opacity of the tree canopy was derived from incoming brightness temperatures received on the ground. A collocated eddy-covariance flux tower measured water vapor fluxes and meteorological variables above the canopy. In addition, xylem sapflow of trees was measured within the flux tower footprint. We observed considerable diurnal differences between measured canopy opacities and modeled theoretical opacities that were closely linked to xylem sapflow. It is speculated that dielectric changes in the leaves induced by the sapflow are causing the observed diurnal changes. In addition, canopy intercepted rain and dew formation also modulated the diurnal opacity cycle. With an enhanced canopy opacity model accounting for water deposited on the leaves, we quantified the influence of canopy stored water (i.e. intercepted water and dew) on the opacity. A time series of dew formation and rain interception was directly monitored during a period of two weeks. We found that during light rainfall up to 60% of the rain amount is intercepted by the canopy whereas during periods of intense rainfall only 4% were intercepted. On average, 0.17 mm of dew was formed during the night. Dew evaporation contributed 5% to the total water vapor flux measured above the canopy.

A new model of net ecosystem carbon exchange for the deciduous-dominated forest by integrating MODIS and flux data

The eddy covariance technique provides continuous measurements of plot-level net ecosystem carbon exchange (NEE) across a wide range of vegetation types. However, these NEE estimates only represent fluxes at the tower footprint scale. To quantify the NEE over regions or continents, flux tower measurements need to be up-scaled to large areas. In the present study, we propose a new NEE model solely based on Moderate Resolution Imaging Spectroradiometer data, including enhanced vegetation index (EVI), land surface water index (LWSI), land surface temperature (LST), and Terra nighttime LST′. Site-specific data from the deciduous-dominated Harvard Forest flux site were used. Analysis covered six years (2001–2006) of CO 2 flux data. The data of the first four years were used for model building and the rest as validation set. Compared with the model based solely on EVI, we also introduced LST and LSWI into the new model. The results showed that this method could further improve the precision ( R 2 and RMSE reached 0.857 and 1.273, respectively) and generally capture the expected seasonal patterns of NEE.

A data-model fusion approach for upscaling gross ecosystem productivity to the landscape scale based on remote sensing and flux footprint modelling

Abstract. In order to use the global available eddy-covariance (EC) flux dataset and remote-sensing measurements to provide estimates of gross primary productivity (GPP) at landscape (101–102 km2), regional (103–106 km2) and global land surface scales, we developed a satellite-based GPP algorithm using LANDSAT data and an upscaling framework. The satellite-based GPP algorithm uses two improved vegetation indices (Enhanced Vegetation Index – EVI, Land Surface Water Index – LSWI). The upscalling framework involves flux footprint climatology modelling and data-model fusion. This approach was first applied to an evergreen coniferous stand in the subtropical monsoon climatic zone of south China. The EC measurements at Qian Yan Zhou tower site (26°44´48" N, 115°04´13" E), which belongs to the China flux network and the LANDSAT and MODIS imagery data for this region in 2004 were used in this study. A consecutive series of LANDSAT-like images of the surface reflectance at an 8-day interval were predicted by blending the LANDSAT and MODIS images using an existing algorithm (ESTARFM: Enhanced Spatial and Temporal Adaptive Reflectance Fusion Model). The seasonal dynamics of GPP were then predicted by the satellite-based algorithm. MODIS products explained 60% of observed variations of GPP and underestimated the measured annual GPP (= 1879 g C m−2) by 25–30%; while the satellite-based algorithm with default static parameters explained 88% of observed variations of GPP but overestimated GPP during the growing seasonal by about 20–25%. The optimization of the satellite-based algorithm using a data-model fusion technique with the assistance of EC flux tower footprint modelling reduced the biases in daily GPP estimations from about 2.24 g C m−2 day−1 (non-optimized, ~43.5% of mean measured daily value) to 1.18 g C m−2 day−1 (optimized, ~22.9% of mean measured daily value). The remotely sensed GPP using the optimized algorithm can explain 92% of the seasonal variations of EC observed GPP. These results demonstrated the potential combination of the satellite-based algorithm, flux footprint modelling and data-model fusion for improving the accuracy of landscape/regional GPP estimation, a key component for the study of the carbon cycle.

Interannual, seasonal, and diel variation in soil respiration relative to ecosystem respiration at a wetland to upland slope at Harvard Forest

Soil carbon dioxide efflux (soil respiration, SR) was measured with eight autochambers at two locations along a wetland to upland slope at Harvard Forest over a 4 year period, 2003–2007. SR was consistently higher in the upland plots than at the wetland margin during the late summer/early fall. Seasonal and diel hystereses with respect to soil temperatures were of sufficient magnitude to prevent quantification of the influence of soil moisture, although apparent short‐term responses of SR to precipitation occurred. Calculations of annual cumulative SR illustrated a decreasing trend in SR over the 5 year period, which were correlated with decreasing springtime mean soil temperatures. Spring soil temperatures decreased despite rising air temperatures over the same period, possibly as an effect of earlier leaf expansion and shading. The synchronous decrease in spring soil temperatures and SR during regional warming of air temperatures may represent a negative feedback on a warming climate by reducing CO2 production from soils. SR reached a maximum later in the year than total ecosystem respiration (ER) measured at a nearby eddy covariance flux tower, and the seasonality of their temperature response patterns were roughly opposite. SR, particularly in the upland, exceeded ER in the late summer/early fall in each year, suggesting that areas of lower efflux such as the wetland may be significant in the flux tower footprint or that long‐term bias in either estimate may create a mismatch. Annual estimates of ER decreased over the same period and were highly correlated with SR.

Fitossociologia de um Cerrado denso em área de influência de torre de fluxo de carbono, Pé-de-Gigante, Parque Estadual de Vassununga, SP

O domínio do Cerrado compreende uma área contínua nos estados centrais do Brasil e áreas disjuntas em outros estados, incluindo São Paulo. Essa vegetação ocupava originalmente 21% do território brasileiro, restando atualmente apenas 21,6% de sua extensão original. A área recoberta por essa vegetação em São Paulo cobria 14% de sua área total e seus remanescentes recobrem menos de 1% da ocorrência original dessa vegetação. Estudos recentes indicam que o valor da produtividade líquida no Cerrado Pé-de-Gigante (SP) constitui um pequeno dreno de carbono e indicou que a sazonalidade foi o fator determinante do valor observado. Os estudos dos fluxos de carbono em ecossistemas terrestres são raramente acompanhados de abordagens ecofisiológicas de modo a explorar a relação funcional das espécies que compõem o ecossistema e os valores líquidos obtidos para o mesmo. Assim, o objetivo deste trabalho foi caracterizar estruturalmente a vegetação presente na área de maior influência da torre de fluxo instalada no Cerrado Pé-de-Gigante, visando possibilitar estudos relacionados à quantificação em longo prazo da dinâmica dos fluxos de água, energia e CO2 na vegetação de Cerrado. Para isso foram levantadas 20 parcelas (10 x 10 m) em 0,2 ha de Cerrado, e amostraram-se todas as plantas com perímetro ao nível do solo >6 cm (exceto lianas e árvores mortas). A distribuição das classes de diâmetro e estrutura vertical, assim como os parâmetros fitossociológicos foram analisados. Encontramos 1451 indivíduos, distribuídos em 85 espécies, 52 gêneros e 31 famílias. A densidade absoluta e área basal foram de 7255 ind. ha-1 e de 7,9 m².ha-1, respectivamente. A família Leguminosae apresentou o maior número de espécies (13). O Índice de diversidade de Shannon (H') foi 3,27 nats.ind-1. A distribuição em classes de diâmetro mostrou uma curva de "J" invertido, estando a maioria dos indivíduos na primeira classe. Concluímos que a área deve ser classificada como Cerrado denso, devido principalmente à dominância pela espécie arbórea Anadenanthera falcata, cuja ocorrência no estado foi relatada apenas em locais com solos ricos em saturação de bases na região das Cuestas Basálticas, devido também à maior área basal dos indivíduos, comparando com outros fragmentos de Cerrado. Além da espécie citada, Myrcia lingua e Xylopia aromatica, apresentaram os maiores IVI (Valor de importância).

A simple filtered photodiode instrument for continuous measurement of narrowband NDVI and PRI over vegetated canopies

Recent advances in understanding relationships between spectral reflectance of vegetation canopies and the structural and physiological drivers of canopy-atmosphere carbon dioxide exchange highlight the potential for using narrowband spectral vegetation indices to spatially scale CO 2 fluxes beyond the area of a tower footprint. However, ground reference observations of narrowband spectral reflectance in support of satellite observations can be challenging to obtain because (1) automated sampling of both upwelling and downwelling radiation is required over extended time periods to characterize diurnal and seasonal variability, (2) hyperspectral spectroradiometer data and hardware can be sensitive to environmental factors such as temperature and moisture, and (3) hyperspectral spectroradiometers are expensive, greatly limiting prospects for widespread automated sampling. We have therefore developed the QuadPod: a simple, lightweight, relatively low cost and low power sensor capable of continuously measuring upwelling and downwelling radiation in 10 nm wavebands centered at 532 nm, 568 nm, 676 nm, and 800 nm. QuadPod measurements can be combined to calculate spectral reflectance indices (e.g., the photochemical reflectance index, PRI; and the normalized difference vegetation index, NDVI) useful for modeling canopy-atmosphere carbon exchange. The basic QuadPod instrument design described here can be implemented using any combination of optical filters in order to calculate other spectral vegetation indices.

A continuous measure of gross primary production for the conterminous United States derived from MODIS and AmeriFlux data

The quantification of carbon fluxes between the terrestrial biosphere and the atmosphere is of scientific importance and also relevant to climate-policy making. Eddy covariance flux towers provide continuous measurements of ecosystem-level exchange of carbon dioxide spanning diurnal, synoptic, seasonal, and interannual time scales. However, these measurements only represent the fluxes at the scale of the tower footprint. Here we used remotely sensed data from the Moderate Resolution Imaging Spectroradiometer (MODIS) to upscale gross primary productivity (GPP) data from eddy covariance flux towers to the continental scale. We first combined GPP and MODIS data for 42 AmeriFlux towers encompassing a wide range of ecosystem and climate types to develop a predictive GPP model using a regression tree approach. The predictive model was trained using observed GPP over the period 2000–2004, and was validated using observed GPP over the period 2005–2006 and leave-one-out cross-validation. Our model predicted GPP fairly well at the site level. We then used the model to estimate GPP for each 1 km × 1 km cell across the U.S. for each 8-day interval over the period from February 2000 to December 2006 using MODIS data. Our GPP estimates provide a spatially and temporally continuous measure of gross primary production for the U.S. that is a highly constrained by eddy covariance flux data. Our study demonstrated that our empirical approach is effective for upscaling eddy flux GPP data to the continental scale and producing continuous GPP estimates across multiple biomes. With these estimates, we then examined the patterns, magnitude, and interannual variability of GPP. We estimated a gross carbon uptake between 6.91 and 7.33 Pg C yr − 1 for the conterminous U.S. Drought, fires, and hurricanes reduced annual GPP at regional scales and could have a significant impact on the U.S. net ecosystem carbon exchange. The sources of the interannual variability of U.S. GPP were dominated by these extreme climate events and disturbances.

Digital image-derived greenness links deep soil moisture to carbon uptake in a creosotebush-dominated shrubland

Changes in the timing, frequency, and magnitude of precipitation events are projected for semiarid ecosystems worldwide. The ecological consequences associated with these precipitation changes will be better understood if the hydrological triggers of vegetation response can be better identified. Previous research has suggested that soil moisture, likely from large monsoon rainstorms, plays a critical role in triggering the phenological response of semiarid shrublands. Here we propose that the recent emergence of time-lapse repeat digital photography (pheno-cams) can play a role in further explaining the hydrological triggers of phenological response in semiarid shrublands. This study is focused on a creosotebush-dominated ecosystem of the Santa Rita Experimental Range, southeastern Arizona. In addition to typical eddy covariance instrumentation, this site offers continuous measurements of soil moisture in 6 one-meter profiles. Additionally, three pheno-cams have been installed in the footprint of the eddy covariance tower at the site. We demonstrate (1) that the green-up of evergreen creosotebush can be tracked using an average greenness index calculated from multiple pheno-cams within a tower footprint; (2) that the green-up of creosotebush is driven by deep soil moisture (e.g. > 30 cm); and (3) that carbon uptake can be predicted using image-derived green-up of creosotebush.

Scale-dependency of surface fluxes in an atmospheric mesoscale model: effect of spatial heterogeneity in atmospheric conditions

Abstract. We examined the nonlinear effect of spatial heterogeneity in atmospheric conditions on the simulation of surface fluxes in the mesoscale model, MM5 by testing their scale-invariance from a tower footprint to regional scales. The test domain was a homogeneous shortgrass prairie in the central part of the Tibetan Plateau with an eddy-covariance flux tower at the center. We found that the spatial variability resulting from changing distribution of clouds and precipitation in the model domain affected radiative forcing at the ground surface, thereby altering the partitioning of surface fluxes. Consequently, due to increasing spatial variability in atmospheric conditions, the results of MM5 did not produce convergent estimates of surface fluxes with increasing grid sizes. Our finding demonstrates that an atmospheric model can underestimate surface fluxes in regional scale not necessarily due to intrinsic model inaccuracy (e.g., inaccurate parameterization) but due to scale-dependent nonlinear effect of spatial variability in atmospheric conditions.

Estimation of net ecosystem carbon exchange for the conterminous United States by combining MODIS and AmeriFlux data

Eddy covariance flux towers provide continuous measurements of net ecosystem carbon exchange (NEE) for a wide range of climate and biome types. However, these measurements only represent the carbon fluxes at the scale of the tower footprint. To quantify the net exchange of carbon dioxide between the terrestrial biosphere and the atmosphere for regions or continents, flux tower measurements need to be extrapolated to these large areas. Here we used remotely sensed data from the Moderate Resolution Imaging Spectrometer (MODIS) instrument on board the National Aeronautics and Space Administration's (NASA) Terra satellite to scale up AmeriFlux NEE measurements to the continental scale. We first combined MODIS and AmeriFlux data for representative U.S. ecosystems to develop a predictive NEE model using a modified regression tree approach. The predictive model was trained and validated using eddy flux NEE data over the periods 2000–2004 and 2005–2006, respectively. We found that the model predicted NEE well ( r = 0.73, p < 0.001). We then applied the model to the continental scale and estimated NEE for each 1 km × 1 km cell across the conterminous U.S. for each 8-day interval in 2005 using spatially explicit MODIS data. The model generally captured the expected spatial and seasonal patterns of NEE as determined from measurements and the literature. Our study demonstrated that our empirical approach is effective for scaling up eddy flux NEE measurements to the continental scale and producing wall-to-wall NEE estimates across multiple biomes. Our estimates may provide an independent dataset from simulations with biogeochemical models and inverse modeling approaches for examining the spatiotemporal patterns of NEE and constraining terrestrial carbon budgets over large areas.

Effect of remote sensing spatial resolution on interpreting tower-based flux observations

Validation comparisons between satellite-based surface energy balance models and tower-based flux measurements over heterogeneous landscapes can be strongly influenced by the spatial resolution of the remote sensing inputs. In this paper, a two-source energy balance model developed to use thermal and visible /near-infrared remotely sensed data is applied to Landsat imagery collected during the 2004 Soil Moisture Experiment (SMEX04) conducted in southern Arizona. Using a two dimensional flux-footprint algorithm, modeled surface fluxes are compared to tower measurements at three locations in the SMEX04 study area: two upland sites, and one riparian site. The effect of pixel resolution on evaluating the performance of the land surface model and interpreting spatial variations of land surface fluxes over these heterogeneous areas is evaluated. Three Landsat scenes were examined, one representing the dry season and the other two representing the relatively wet monsoon season. The model was run at three resolution scales: namely the Landsat visible/near-infrared band resolution (30 m), the Landsat 5 thermal band resolution (120 m), and 960 m, which is nominally the MODIS thermal resolution at near-nadir. Comparisons between modeled and measured fluxes at the three tower sites showed good agreement at the 30 m and 120 m resolutions — pixel scales at which the source area influencing the tower measurement (∼ 100 m) is reasonably resolved. At 960 m, the agreement is relatively poor, especially for the latent heat flux, due to sub-pixel heterogeneity in land surface conditions at scales exceeding the tower footprint. Therefore in this particular landscape, thermal data at 1-km resolution are not useful in assessing the intrinsic accuracy of the land-surface model in comparison with tower fluxes. Furthermore, important spatial patterns in the landscape are lost at this resolution. Currently, there are no definite plans supporting high resolution thermal data with regular global coverage below ∼ 700 m after Landsat 5 and ASTER fail. This will be a serious problem for the application and validation of thermal-based land-surface models over heterogeneous landscapes.

Spatial modelling of the fraction of photosynthetically active radiation absorbed by a boreal mixedwood forest using a lidar–hyperspectral approach

The spatial variability of the fraction of photosynthetically active radiation absorbed by the canopy (fPAR) was characterized for a heterogeneous boreal mixedwood forest site located in northern Ontario, Canada, based on relationships found between fPAR and light detection and ranging (lidar) data over different canopy architectures. Estimates of fPAR were derived from radiation measurements made above the canopy at a flux tower and below-canopy radiation was measured across a range of species compositions and canopy architectures. Airborne lidar data were used to characterize spatial variability of canopy structure around the flux tower and a map of mean canopy chlorophyll concentration was derived from airborne hyperspectral imagery. Different volumes of lidar points for the locations directly above each photosynthetically active radiation (PAR) sensor were examined to determine if there is an optimal method of relating lidar returns to estimated fPAR values. The strongest correlations between mean lidar height and fPAR occurred when using points that fell within a theoretical cone which originated at the PAR sensor having a solid angle α = 55°. For diffuse conditions, the correlation ( r) between mean lidar height versus fPAR × chlorophyll was stronger than between mean lidar height versus fPAR by 8% for mean daily fPAR and from 10 to 20% for diurnal fPAR, depending on solar zenith angle. For direct light conditions, the relationship was improved by 12% for mean daily fPAR and 12–41% for diurnal relationships. Linear regression models of mean daily fPAR × chlorophyll versus mean lidar height were used in conjunction with gridded lidar data and the canopy chlorophyll map to generate maps of mean daily fPAR for direct and diffuse sunlight conditions. Site average fPAR calculated from these maps was 0.79 for direct light conditions and 0.78 for diffuse conditions. When compared to point estimates of mean daily fPAR calculated on the tower, the average fPAR was significantly lower than the point estimate. Subtracting the direct sunlight fPAR map from the diffuse sunlight fPAR map revealed a distinct spatial pattern showing that areas with open canopies and relatively low chlorophyll (e.g., black spruce patches) have a higher fPAR under direct sunlight conditions, while closed canopies with higher chlorophyll (e.g., deciduous species) absorb more PAR under diffuse conditions. These findings have implications for scaling from point measurements at flux towers to larger resolution satellite imagery and addressing local scale heterogeneity in flux tower footprints.

Upscaling fluxes from tower to landscape: Overlaying flux footprints on high-resolution (IKONOS) images of vegetation cover

In this paper, we describe the process of assessing tower footprint climatology, spatial variability of site vegetation density based on satellite image analysis, and sensor location bias in scaling up to 1 km × 1 km patch. Three flat sites with different vegetation cover and surface heterogeneity were selected from AmeriFlux tower sites: the oak/grass site and the annual grassland site in a savannah ecosystem in northern California and a slash pine forest site in Florida, USA. The site vegetation density was expressed in terms of normalized difference vegetation index (NDVI) and crown closure (CC) by analyzing the high-resolution IKONOS satellite image. At each site, the spatial structure of vegetation density was characterized using semivariogram and window size analyses. Footprint maps were produced by a simple model based on the analytical solution of the Eulerian advection–diffusion equation. The resulting horizontal arrays of footprint functions were then superimposed with those of NDVI and CC. Annual sensor location biases for the oak/grass and the pine forest sites were <4% for both NDVI and CC, requiring no flux corrections in scaling from tower to landscape of 1 km 2. Although the annual grassland site displayed much larger location biases (28% for NDVI, 94% for CC), their temporal changes associated with averaging time showed a real potential to develop algorithms aimed at upscaling tower fluxes to the landscape in an effort to provide validation data for MODIS products.