Yanlian Zhou

Water use efficiency of China’s terrestrial ecosystems and responses to drought

Water use efficiency (WUE) measures the trade-off between carbon gain and water loss of terrestrial ecosystems, and better understanding its dynamics and controlling factors is essential for predicting ecosystem responses to climate change. We assessed the magnitude, spatial patterns, and trends of WUE of China’s terrestrial ecosystems and its responses to drought using a process-based ecosystem model. During the period from 2000 to 2011, the national average annual WUE (net primary productivity (NPP)/evapotranspiration (ET)) of China was 0.79 g C kg−1 H2O. Annual WUE decreased in the southern regions because of the decrease in NPP and the increase in ET and increased in most northern regions mainly because of the increase in NPP. Droughts usually increased annual WUE in Northeast China and central Inner Mongolia but decreased annual WUE in central China. “Turning-points” were observed for southern China where moderate and extreme droughts reduced annual WUE and severe drought slightly increased annual WUE. The cumulative lagged effect of drought on monthly WUE varied by region. Our findings have implications for ecosystem management and climate policy making. WUE is expected to continue to change under future climate change particularly as drought is projected to increase in both frequency and severity.

Recent trends in vegetation greenness in China significantly altered annual evapotranspiration and water yield

There has been growing evidence that vegetation greenness has been increasing inmany parts of the northernmiddle and high latitudes includingChina during the last three to four decades. However, the effects of increasing vegetation greenness particularly afforestation on the hydrological cycle have been controversial.We used a process-based ecosystemmodel and a satellite-derived leaf area index (LAI) dataset to examine how the changes in vegetation greenness affected annual evapotranspiration (ET) andwater yield for China over the period from2000 to 2014. Significant trends in vegetation greenness were observed in 26.1%ofChina’s land area.We used twomodel simulations drivenwith original and detrended LAI, respectively, to assess the effects of vegetation ‘greening’ and ‘browning’ on terrestrial ET andwater yield. On a per-pixel basis, vegetation greening increased annual ET and decreasedwater yield, while vegetation browning reduced ET and increasedwater yield. At the large river basin and national scales, the greening trends also had positive effects on annual ET and had negative effects onwater yield. Our results showed that the effects of the changes in vegetation greenness on the hydrological cycle variedwith spatial scale. Afforestation efforts perhaps should focus on southernChinawith larger water supply given thewater crisis in northernChina and the negative effects of vegetation greening onwater yield. Future studies on the effects of the greenness changes on the hydrological cycle are needed to account for the feedbacks to the climate.

Global parameterization and validation of a two-leaf light use efficiency model for predicting gross primary production across FLUXNET sites

Light use efficiency (LUE) models are widely used to simulate gross primary production (GPP). However, the treatment of the plant canopy as a big leaf by these models can introduce large uncertainties in simulated GPP. Recently, a two-leaf light use efficiency (TL-LUE) model was developed to simulate GPP separately for sunlit and shaded leaves and has been shown to outperform the big-leaf MOD17 model at six FLUX sites in China. In this study we investigated the performance of the TL-LUE model for a wider range of biomes. For this we optimized the parameters and tested the TL-LUE model using data from 98 FLUXNET sites which are distributed across the globe. The results showed that the TL-LUE model performed in general better than the MOD17 model in simulating 8 day GPP. Optimized maximum light use efficiency of shaded leaves (epsilon(msh)) was 2.63 to 4.59 times that of sunlit leaves (epsilon(msu)). Generally, the relationships of epsilon(msh) and epsilon(msu) with epsilon(max) were well described by linear equations, indicating the existence of general patterns across biomes. GPP simulated by the TL-LUE model was much less sensitive to biases in the photosynthetically active radiation (PAR) input than the MOD17 model. The results of this study suggest that the proposed TL-LUE model has the potential for simulating regional and global GPP of terrestrial ecosystems, and it is more robust with regard to usual biases in input data than existing approaches which neglect the bimodal within-canopy distribution of PAR.

Prediction of summer grain crop yield with a process-based ecosystem model and remote sensing data for the northern area of the Jiangsu Province, China

Yield prediction is important for agricultural management, food security warning and food trade policy. Remote sensing has been a useful tool for predicting crop yields. In this study, a modified daily process-based ecosystem model (the Boreal Ecosystem Productivity Simulator) is employed in conjunction with land cover and leaf area index (LAI) products from MODIS to predict summer grain crop yields in the northern area of the Yangtze River in the Jiangsu Province, China. The model was driven by soil texture, land cover, daily meteorological and MODIS LAI data for 2004–2006. Simulated growing season net primary productivity (NPP) of summer grain crops (November–May) and census data of crop yields in 2004 were used to derive the county-level harvest index, which is then used in conjunction with simulated NPP to predict crop yields in 2005 and 2006. The model captures 89 % and 88 % of variations in crop yields at county-level compared with census data in 2005 and 2006, respectively. The root mean square errors are 265 and 277 kg ha−1 in these two years. The results show the usefulness of a process-based model driven by remote sensing in predicting crop yields. In such predictions, the considerable spatial variability of the harvest index should be taken into consideration.

Evaluation and improvement of MODIS gross primary productivity in typical forest ecosystems of East Asia based on eddy covariance measurements

Gross primary productivity (GPP) is a major component of carbon exchange between the atmosphere and terrestrial ecosystems and a key component of the terrestrial carbon cycle. Because of the large spatial heterogeneity and temporal dynamics of ecosystems, it is a challenge to estimate GPP accurately at global or regional scales. The 8-day MODerate resolution Imaging Spectroradiometer (MODIS) GPP product provides a near real time estimate of global GPP. However, previous studies indicated that MODIS GPP has large uncertainties, partly caused by biases in parameterization and forcing data. In this study, MODIS GPP was validated using GPP derived from the eddy covariance flux measurements at five typical forest sites in East Asia. The validation indicated that MODIS GPP was seriously underestimated in these forest ecosystems of East Asia, especially at northern sites. With observed meteorological data, fraction of photosynthetically active radiation absorbed by the plant canopy (fPAR) calculated using smoothed MODIS leaf area index, and optimized maximum light use efficiency (εmax) to force the MOD17 algorithm, the agreement between predicted GPP and tower-based GPP was significantly improved. The errors of MODIS GPP in these forest ecosystems of East Asia were mainly caused by uncertainties in εmax, followed by those in fPAR and meteorological data. The separation of canopy into sunlit and shaded leaves, for which GPP is individually calculated, can improve GPP simulation significantly.

Estimation of forest height, biomass and volume using support vector regression and segmentation from lidar transects and Quickbird imagery

Lidar (light detection and ranging) remote sensing can accurately characterize forest vertical structure, such as canopy height, above-ground biomass (AGB) and timber volume; however, data acquisition is expensive. To reduce costs, one potential method is to integrate (small area) lidar transects and (large extent) optical imagery to estimate forest characteristics. Typically, multiple regression is used to link variables extracted from lidar transect data and optical imagery. Height information is then generalized from the area covered by lidar transects to other areas without lidar coverage. However, multiple regression models may not fully capture the complex relationship between variables. Fortunately, Support vector regression (SVR) provides a solution to deal with such complex nonlinear problems. Using a case study in Vancouver Island, Canada, SVR was applied to generalize canopy height from lidar transect(s) to the entire study area (2601 ha) based on a segmented Quickbird image. Results show that: (i) compared to typical multiple regression models, the SVR models provided better results for estimating canopy height; (ii) by using only one lidar transect (i.e., 8.8% cover), the SVR model generates an average canopy height estimation error of 6.2 m — which is less than a British Columbia forest inventory height class (9.0 m); and (iii) the final model estimates have relatively high correlations with field data for forest canopy height (R2: 0.81), AGB (R2: 0.76) and volume (R2: 0.64), while representing dramatically reduced acquisition costs.

Performance of Linear and Nonlinear Two-Leaf Light Use Efficiency Models at Different Temporal Scales

The reliable simulation of gross primary productivity (GPP) at various spatial and temporal scales is of significance to quantifying the net exchange of carbon between terrestrial ecosystems and the atmosphere. This study aimed to verify the ability of a nonlinear two-leaf model (TL-LUEn), a linear two-leaf model (TL-LUE), and a big-leaf light use efficiency model (MOD17) to simulate GPP at half-hourly, daily and 8-day scales using GPP derived from 58 eddy-covariance flux sites in Asia, Europe and North America as benchmarks. Model evaluation showed that the overall performance of TL-LUEn was slightly but not significantly better than TL-LUE at half-hourly and daily scale, while the overall performance of both TL-LUEn and TL-LUE were significantly better (p < 0.0001) than MOD17 at the two temporal scales. The improvement of TL-LUEn over TL-LUE was relatively small in comparison with the improvement of TL-LUE over MOD17. However, the differences between TL-LUEn and MOD17, and TL-LUE and MOD17 became less distinct at the 8-day scale. As for different vegetation types, TL-LUEn and TL-LUE performed better than MOD17 for all vegetation types except crops at the half-hourly scale. At the daily and 8-day scales, both TL-LUEn and TL-LUE outperformed MOD17 for forests. However, TL-LUEn had a mixed performance for the three non-forest types while TL-LUE outperformed MOD17 slightly for all these non-forest types at daily and 8-day scales. The better performance of TL-LUEn and TL-LUE for forests was mainly achieved by the correction of the underestimation/overestimation of GPP simulated by MOD17 under low/high solar radiation and sky clearness conditions. TL-LUEn is more applicable at individual sites at the half-hourly scale while TL-LUE could be regionally used at half-hourly, daily and 8-day scales. MOD17 is also an applicable option regionally at the 8-day scale.

Close relationship between spectral vegetation indices and V-cmax in deciduous and mixed forests

Seasonal variations of photosynthetic capacity parameters, notably the maximum carboxylation rate, Vcmax, play an important role in accurate estimation of CO2 assimilation in gas-exchange models. Satellite-derived normalised difference vegetation index (NDVI), enhanced vegetation index (EVI) and model-data fusion can provide means to predict seasonal variation in Vcmax. In this study, Vcmax was obtained from a process-based model inversion, based on an ensemble Kalman filter (EnKF), and gross primary productivity, and sensible and latent heat fluxes measured using eddy covariance technique at two deciduous broadleaf forest sites and a mixed forest site. Optimised Vcmax showed considerable seasonal and inter-annual variations in both mixed and deciduous forest ecosystems. There was noticeable seasonal hysteresis in Vcmax in relation to EVI and NDVI from 8 d composites of satellite data during the growing period. When the growing period was phenologically divided into two phases (increasing VIs and decreasing VIs phases), significant seasonal correlations were found between Vcmax and VIs, mostly showing R2>0.95. Vcmax varied exponentially with increasing VIs during the first phase (increasing VIs), but second and third-order polynomials provided the best fits of Vcmax to VIs in the second phase (decreasing VIs). The relationships between NDVI and EVI with Vcmax were different. Further efforts are needed to investigate Vcmax–VIs relationships at more ecosystem sites to the use of satellite-based VIs for estimating Vcmax.

Significant Decrease of Uncertainties in Sensible Heat Flux Simulation Using Temporally Variable Aerodynamic Roughness in Two Typical Forest Ecosystems of China

AbstractAerodynamic roughness length zom is an important parameter for reliably simulating surface fluxes. It varies with wind speed, atmospheric stratification, terrain, and other factors. However, it is usually considered a constant. It is known that uncertainties in zom result in latent heat flux (LE) simulation errors, since zom links LE with aerodynamic resistance. The effects of zom on sensible heat flux (SH) simulation are usually neglected because there is no direct link between the two. By comparing SH simulations with three types of zom inputs, it is found that allowing zom temporal variation in an SH simulation model significantly improves agreement between simulated and measured SH and also decreases the sensitivity of the SH model to the heat transfer coefficient Ct, which in turn determines the linkage between zom and thermal roughness length zoh.

Ability of the Photochemical Reflectance Index to Track Light Use Efficiency for a Sub-Tropical Planted Coniferous Forest

Light use efficiency (LUE) models are widely used to estimate gross primary productivity (GPP), a dominant component of the terrestrial carbon cycle. Their outputs are very sensitive to LUE. Proper determination of this parameter is a prerequisite for LUE models to simulate GPP at regional and global scales. This study was devoted to investigating the ability of the photochemical reflectance index (PRI) to track LUE variations for a sub-tropical planted coniferous forest in southern China using tower-based PRI and GPP measurements over the period from day 101 to 275 in 2013. Both half-hourly PRI and LUE exhibited detectable diurnal and seasonal variations, and decreased with increases of vapor pressure deficit (VPD), air temperature (Ta), and photosynthetically active radiation (PAR). Generally, PRI is able to capture diurnal and seasonal changes in LUE. However, correlations of PRI with LUE varied dramatically throughout the growing season. The correlation was the strongest (R2 = 0.6427, p 0.3) with moderate to high VPD (>20 hPa) and high temperatures (>31 C). Overall, we found that PRI is most sensitive to variations in LUE under stressed conditions, and the sensitivity decreases as the growing conditions become favorable when atmosphere water vapor, temperature and soil moisture are near the optimum conditions.