Ground Plots Research Articles

Mapping Annual Forest Cover in Sub-Humid and Semi-Arid Regions through Analysis of Landsat and PALSAR Imagery

Accurately mapping the spatial distribution of forests in sub-humid to semi-arid regions over time is important for forest management but a challenging task. Relatively large uncertainties still exist in the spatial distribution of forests and forest changes in the sub-humid and semi-arid regions. Numerous publications have used either optical or synthetic aperture radar (SAR) remote sensing imagery, but the resultant forest cover maps often have large errors. In this study, we propose a pixel- and rule-based algorithm to identify and map annual forests from 2007 to 2010 in Oklahoma, USA, a transitional region with various climates and landscapes, using the integration of the L-band Advanced Land Observation Satellite (ALOS) PALSAR Fine Beam Dual Polarization (FBD) mosaic dataset and Landsat images. The overall accuracy and Kappa coefficient of the PALSAR/Landsat forest map were about 88.2% and 0.75 in 2010, with the user and producer accuracy about 93.4% and 75.7%, based on the 3270 random ground plots collected in 2012 and 2013. Compared with the forest products from Japan Aerospace Exploration Agency (JAXA), National Land Cover Database (NLCD), Oklahoma Ecological Systems Map (OKESM) and Oklahoma Forest Resource Assessment (OKFRA), the PALSAR/Landsat forest map showed great improvement. The area of the PALSAR/Landsat forest was about 40,149 km2 in 2010, which was close to the area from OKFRA (40,468 km2), but much larger than those from JAXA (32,403 km2) and NLCD (37,628 km2). We analyzed annual forest cover dynamics, and the results show extensive forest cover loss (2761 km2, 6.9% of the total forest area in 2010) and gain (3630 km2, 9.0%) in southeast and central Oklahoma, and the total area of forests increased by 684 km2 from 2007 to 2010. This study clearly demonstrates the potential of data fusion between PALSAR and Landsat images for mapping annual forest cover dynamics in sub-humid to semi-arid regions, and the resultant forest maps would be helpful to forest management.

A Simulation of Image-Assisted Forest Monitoring for National Inventories

The efficiency of national forest monitoring efforts can be increased by the judicious incorporation of ancillary data. For instance, a fixed number of ground plots might be used to inform a larger set of annual estimates by observing a smaller proportion of the plots each year while augmenting each annual estimate with ancillary data in order to reduce overall costs while maintaining a desired level of accuracy. Differencing successive geo-rectified remotely sensed images can conceivably provide forest change estimates at a scale and level of accuracy conducive to the improvement of temporally relevant forest attribute estimates. Naturally, the degree of improvement in the desired estimates is highly dependent on the relationships between the spatial-temporal scales of ground plot and remotely sensed observations and the desired spatial-temporal scale of estimation. In this paper, fixed scales of observation for each data source are used to explore the value of three different levels of information available from the remotely sensed image-change estimates. Four populations are simulated and sampled under four sampling error structures. The results show that the image change estimates (ICE) can be used to significantly reduce bias for annual estimates of harvest and mortality and that improved estimation of harvest and mortality can sometimes, but not always, contribute to better estimates of standing volume.

Model-based estimation of above-ground biomass in the miombo ecoregion of Zambia

Information on above-ground biomass (AGB) is important for managing forest resource use at local levels, land management planning at regional levels, and carbon emissions reporting at national and international levels. In many tropical developing countries, this information may be unreliable or at a scale too coarse for use at local levels. There is a vital need to provide estimates of AGB with quantifiable uncertainty that can facilitate land use management and policy development improvements. Model-based methods provide an efficient framework to estimate AGB. Using National Forest Inventory (NFI) data for a ~1,000,000 ha study area in the miombo ecoregion, Zambia, we estimated AGB using predicted canopy cover, environmental data, disturbance data, and Landsat 8 OLI satellite imagery. We assessed different combinations of these datasets using three models, a semiparametric generalized additive model (GAM) and two nonlinear models (sigmoidal and exponential), employing a genetic algorithm for variable selection that minimized root mean square prediction error (RMSPE), calculated through cross-validation. We compared model fit statistics to a null model as a baseline estimation method. Using bootstrap resampling methods, we calculated 95 % confidence intervals for each model and compared results to a simple estimate of mean AGB from the NFI ground plot data. Canopy cover, soil moisture, and vegetation indices were consistently selected as predictor variables. The sigmoidal model and the GAM performed similarly; for both models the RMSPE was ~36.8 tonnes per hectare (i.e., 57 % of the mean). However, the sigmoidal model was approximately 30 % more efficient than the GAM, assessed using bootstrapped variance estimates relative to a null model. After selecting the sigmoidal model, we estimated total AGB for the study area at 64,526,209 tonnes (+/− 477,730), with a confidence interval 20 times more precise than a simple design-based estimate. Our findings demonstrate that NFI data may be combined with freely available satellite imagery and soils data to estimate total AGB with quantifiable uncertainty, while also providing spatially explicit AGB maps useful for management, planning, and reporting purposes.

Two-Step Constrained Nonlinear Spectral Mixture Analysis Method for Mitigating the Collinearity Effect

Spectral mixture analysis (SMA) is widely used to quantify the fraction of each component (endmember) of mixed pixels that contain spectral signals from more than one land surface type. Generally, nonlinear SMA (NSMA) outperforms linear SMA (LSMA) in the vegetation (tree, shrub, crop, and grass) and soil mixture case because NSMA considers the significant multiple scattering that exists for these mixtures. However, compared to LSMA, the bilinear NSMA method, which is a typical physical-based NSMA method, is undermined by its susceptibility to the collinearity effect. In this paper, a two-step constrained NSMA method (referred to as TsC-NSMA) is proposed to mitigate the collinearity effect in the bilinear NSMA method. The theoretical maximum likelihood range is mathematically derived for each endmember fraction, and the ranges are used as additional constraints for the bilinear NSMA method to optimize the unmixing results. Three different data sets, including simulated spectral data, an in situ ground plot spectral measurement, and a Landsat8 Operational Land Imager image, were used to assess the performance of the TsC-NSMA method. The results indicated that TsC-NSMA achieved the highest estimation accuracy for all mixed scenarios which either contain severe endmember collinearity or high noise levels, thereby suggesting its ability to mitigate the collinearity effect in the bilinear NSMA method with the potential to improve the estimation of endmember fractions in practical applications.

Using forest structure to predict the distribution of treed boreal peatlands in Canada

Mapping peatland extent in Canada would contribute important information concerning carbon balance and hydrology. While such mapping, based on air photo interpretation and remote sensing data, has recently improved, maps have been limited to 1:1million scale. We hypothesized that forest structure information from forest inventory plots could be used to predict the presence of forested and treed peatlands in boreal Canada at the ground plot-level, and that a resulting model could be used to predict the distribution of forested and treed peatlands across Canada. Inventory ground plots from the Canadian National Forest Inventory (NFI) with organic soil depth measurements were used to create a model of the presence of treed to forested (canopy cover ranging from sparse to closed) peatlands (greater than 40cm organic soil depth) in boreal Canada. The presence of black spruce (Picea mariana) or larch (Larix laricina), in combination with low stand height and stand age greater than 75years, were the strongest predictors of the presence of peatlands. Bioclimatic variables related to high diurnal and annual temperature variation, consistent with a continental climate, also contributed to the increased predicted presence of treed peatlands. Both logistic and boosted regression tree models showed similar results, with ∼87% accuracy in the discrimination of treed peatlands when validated against an independent set of ground plots. The boosted regression tree model was propagated across Canada using forest attribute raster data layers at 250m resolution from the NFI along with bioclimatic layers. Estimates of treed peatland extent agreed with data points from peat cores with 85–95% accuracy in the Boreal Shield ecozone, although prediction was less accurate in the more southern boreal and Great Lakes forest areas. The resulting map can be used as an input to forest carbon modelling, and the improved knowledge of treed peatland extent will be useful in modelling wildfire or peatland drainage.

Mapping and estimating forest area and aboveground biomass in miombo woodlands in Tanzania using data from airborne laser scanning, TanDEM-X, RapidEye, and global forest maps: A comparison of estimated precision

Abstract Field surveys are often a primary source of data for aboveground biomass (AGB) and forest area estimates — two fundamental parameters in forest resource assessments and for measurement, reporting, and verification (MRV) under the United Nations Collaborative Program on Reducing Emissions from Deforestation and Forest Degradation in Developing Countries (REDD +). However, plot-based estimates of such parameters are often not sufficiently precise for their intended purposes, and especially so in developing and tropical countries in which implementation of extensive sample surveys can be cost-prohibitive or infeasible due to inaccessibility. Remotely sensed data can improve the precision of estimates and thereby reduce the need for field samples. To guide investment decision in MRV systems, comparative analyses of the contribution of different types of remotely sensed data to improve precision of estimates are required. The aim of the current study was to quantify the contribution of data from (1) airborne laser scanning (ALS), (2) interferometric synthetic aperture radar (InSAR) derived from TanDEM-X, (3) RapidEye optical imagery, and global forest map products derived from (4) Landsat and (5) ALOS PALSAR L-band radar imagery to improve precision of AGB and forest area estimates beyond the precision that could be obtained by a pure field-based survey in miombo woodlands of Tanzania. Miombo woodlands is one among the most wide-spread vegetation types in eastern, central, and southern Africa, occupying about 9% of the entire African land area. A 365.6 km2 region in Liwale district in Tanzania served as area of interest for this study. Eighty-eight ground plots distributed on 11 clusters of eight plots each according to a probability-based single-stage cluster sampling design served as field data for regression model calibration used for mapping and estimation of AGB and forest area. Model-assisted estimators were used in the estimation. The relative efficiency (RE) of the ALS-assisted estimates of mean AGB per hectare (variance of the field-based estimate relative to the variance of the ALS-assisted estimate) was 3.6. Relative efficiency translates directly to the factor by which the sample size used for the ALS-assisted estimate would have to be multiplied to arrive at the same precision for a pure field-based estimate. RE values for InSAR and RapidEye were 2.8 and 3.3, while the global Landsat and PALSAR map products contributed only marginally to improve precision (RE = 1.3–1.4). For forest area estimation, ALS-assisted estimates showed an RE of 3.7–4.6, while InSAR, RapidEye, and global Landsat and PALSAR maps resulted in RE values of 1.0–1.3, 2.0–2.1, 1.4–1.8, and 1.7, respectively.

Urban forests and social inequality in the Pacific Northwest

Research has shown there is a positive relationship between urban greenness and the well-being of city residents. But greenness is often unevenly distributed across a city, raising environmental justice issues. In 2011 and 2012 the USDA Forest Service, Forest Inventory and Analysis program installed ground plots in the urbanized areas of Oregon and Washington. We analyze these data for the urban areas west of the Cascade Mountains, linking it with demographic data from the U.S. Census to examine the relationship between greenness and socioeconomic status at a sub-regional scale. To explore some relations between urban forest measures and socioeconomic conditions and measures we developed four models: presence of tree canopy cover with a logistic mixed model, and on a subset of the data, percent tree canopy cover with a linear mixed model and tree count and tree species count with Poisson mixed models. We found that median household income, house value, land use, and years in the Tree City USA program contributed to explaining measures of greenness, such as canopy cover presence, percent canopy cover, tree counts, and tree species counts. This agrees with other studies, but does so at a broad scale covering the most densely populated areas in the Pacific Northwest.

Integrating lidar with Landsat data for subalpine temperate forest aboveground carbon estimation

Accurate forest carbon accounting forms a basis for promoting the development of ecosystem service markets including forest carbon sinks. However, carbon assessments over large forest areas are challenging. Difficulties are compounded by the lack of adequate field observations especially in mountainous regions. In this study, we describe the development of a two-phase sampling framework to evaluate regional aboveground carbon density (ACD) of subalpine temperate forests in northwestern China that includes integrating ground plots, airborne lidar metrics, and Landsat images. During the first phase, an accurate, lidar-derived, ACD inventory network of a representative forested zone (Dayekou Basin) was established on the basis of a modified allometric model by adding crown coverage (CC) as a supplementary variable; cross-validated R2 was 0.88 and root mean square error (RMSE) was 14.7 Mg C ha−1. The outcomes of this step enabled the extension of quasi-field plots required for the representative carbon evaluations and the amplification of the range of observed values. Further integration of lidar measures and optical Landsat data by using the partial least squares regression (PLSR) method was conducted in the subsequent phase. The final model developed for broad-scale estimates explained 76% of the variance in forest ACD and produced a mean bias error of 27.9 Mg C ha−1. Aboveground carbon stocks for the whole ecoregion averaged 77.2 Mg ha−1, which generated an uncertainty of 13%. Visual patterns revealed a systematic overestimation for low ACD values and an underestimation in those regions with high carbon density. Potential errors in our carbon estimates could be associated with the saturation of optical signals, accuracy of land-cover map, and effects of topographic conditions. Overall, the double-sampling method demonstrated promising means for carbon accounting over large areas in a spatially-explicit manner and provided a good first approximation of carbon quantities for the forests in the ecoregion. Our study illustrated the potential for the use of lidar sampling in facilitating scaling of field surveys to a larger spatial extent than ground-based practices by supplying accurate biophysical measurements (e.g. heights).

Evaluating the impact of leaf-on and leaf-off airborne laser scanning data on the estimation of forest inventory attributes with the area-based approach

In this study, we explored the consequences of using leaf-on and leaf-off airborne laser scanning (ALS) data on area-based model outcomes in a lodgepole pine (Pinus contorta var. latifolia Engelm.) dominated forest in the foothills of the Rocky Mountains in Alberta, Canada. We considered eight forest attributes: top height, mean height, Lorey’s mean height, basal area, quadratic mean diameter, merchantable volume, total volume, and total aboveground biomass. We used 787 ground plots for model development, stratified by ALS acquisition conditions (leaf-on or leaf-off) and dominant forest type (coniferous or deciduous). We also generated pooled models that combined leaf-on and leaf-off ALS data and generic models that combined plot data for all forest types. We evaluated differences in ALS metrics and leaf-on and leaf-off model outcomes, as well as the impacts of pooling leaf-on and leaf-off ALS data, creating generic models, and of applying leaf-on models to leaf-off data (and vice versa). In general, leaf-off and leaf-on ALS metrics were not significantly different (p < 0.05), except for the 5th percentile of height (coniferous) and canopy density metrics (deciduous). Overall, coniferous leaf-on and leaf-off models were comparable, with differences in relative root mean square error (RMSE) and bias of <2% for all attributes except volume, which differed by <4%. RMSE and bias for deciduous leaf-on and leaf-off models for height attributes and quadratic mean diameter differed by <2%, whereas models for volume and biomass differed by <7%. These results affirm that leaf-off data can be used in an area-based approach to estimate forest attributes for both coniferous and deciduous forest types. Relative RMSE and bias for pooled models (combining leaf-on and leaf-off ALS data) differed by <2% relative to leaf-on and leaf-off models, suggesting that in the forests studied herein, combining leaf-on and leaf-off data in an area-based approach does not adversely impact model outcomes. Generic models that did not account for forest type had large errors for volume and biomass (e.g., the relative RMSE for merchantable volume was twice as large as forest type specific models). Likewise, the mixing of leaf-on models with leaf-off data and vice versa resulted in large RMSE and bias for both forest types, and therefore mixing of models and data types should be avoided.

A monitoring indicator system for remote sensing of grassland vegetation growth and suitability evaluation – a case study of the Xilingol Grassland in Inner Mongolia, China

Grassland vegetation growth directly reflects plant growth conditions, and growth processes are an important component of ecological status assessments of grasslands and can provide timely guidance for agricultural production. In this study, the Xilingol Grassland in Inner Mongolia was used as the study area, and a monitoring indicator system for the remote sensing of grassland vegetation growth was established based on 16 days of Moderate Resolution Imaging Spectroradiometer (MODIS) data and ground sampling data, which were used to assess the suitability of the indicator system. A monitoring indicator system for the remote sensing of grassland vegetation that included the modified growth index (MGI)-normalized difference vegetation index (NDVI), MGI-enhanced vegetation index (EVI), growth index (GI)-NDVI, and GI-EVI was established by using the year 2000 as the base year, two vegetation indices and difference and normalized difference methods. A model for estimating the ground growth (g) was then constructed by using expert opinion scoring and the Analytic Hierarchy Process (AHP) to determine the weights of vegetation coverage (c), height (h), and yield (y) in the ground plots, with the model calculated as follows: g = 0.2543c + 0.1848 h + 0.5609y. Additionally, the ground growth value was calculated according to the ground growth model, and the values obtained from the remote-sensing indicators in the corresponding region were subjected to a correlation analysis based on this partition. The remote-sensing growth indices suitable for temperate steppe, meadow steppe, and desert steppe regions were GI-EVI, MGI-NDVI, and GI-NDVI, respectively. Finally, vegetation growth in the Xilingol Grassland was evaluated using the optimal remote-sensing GI for each area, and the results indicated that the areas with the greatest growth improvement occurred in the temperate steppe region followed by the meadow steppe region, whereas vegetation growth improvement was insignificant in the desert steppe region. The results of this study have important significance for the economic development and ecological environmental improvement of pastoral and semi-pastoral regions and can be used to determine the optimal use and scientific management of grasslands.

Effects of plant diversity and structural complexity on parasitoid behaviour in a field experiment

1. In natural ecosystems, plants containing hosts for parasitoids are often embedded within heterogeneous plant communities. These plant communities surrounding host‐infested plants may influence the host‐finding ability of parasitoids.2. A release‐recapture‐approach was used to examine whether the diversity and structural complexity of the community surrounding a host‐infested plant influences the aggregation behaviour of the leaf‐miner parasitoidDacnusa sibiricaTelenga and naturally occurring local leaf‐miner parasitoids. Released and locally present parasitoids were collected on pottedJacobaea vulgarisGaertn.plants infested with the generalist leaf‐minerChromatomyia syngenesiaeHardy. The plants were placed in experimentally established plant communities differing in plant diversity (1–9 species) and habitat complexity (bare ground, mown vegetation, and tall vegetation). Additionally, parasitoids were reared out from host mines on the trap plants.3. Plant diversity did not influence the mean number of recapturedD. sibiricaor captures of other locally present parasitoids but the number of recaptured parasitoids was influenced by habitat complexity. NoD. sibiricaparasitoids were recaptured in the bare ground plots or plots with mown vegetation. The mean number of recapturedD. sibiricagenerally increased with increasing complexity of the plant community, whereas locally present parasitoids were captured more frequently in communities with more bare ground. There was a unimodal relationship between the number of reared out parasitoids and diversity of the surrounding vegetation with the highest density of emerged parasitoids at intermediate diversity levels.4. The present study adds to the thus far limited body of literature examining the aggregation behaviour of parasitoids in the field and suggests that the preference of parasitoids to aggregate in complex versus simple vegetation is association specific and thus depends on the parasitoid species as well as the identity of the plant community.

Calibrated tree counting on remotely sensed images of planted forests

The development of robust and accurate methods for counting trees from remotely sensed data could provide substantial cost savings in forest inventory. A new methodology that provides a framework for calibrating tree detection algorithms to obtain accurate tree counts for even-aged stands is described. The methodology was evaluated using two tree detection algorithms and two operators using airborne laser scanning (ALS) and orthophotograph images for four Pinus radiata D.Don stands ranging in age between 5 and 32 years with stand densities ranging between 204 and 826 stems ha−1. For application of the methodology to ALS images the error of estimate on the total count was 4.7% when calibration counts from actual ground plots were used and 10.5% when calibration counts from virtual plots on the image were used. For orthophotographs, the error of estimate was 6.1% using ground calibration plots and 24.3% using calibration counts from virtual plots. The described methodology was shown to be robust to variations in the process from the two operators and two algorithms evaluated. The measure of accuracy determined using the methodology can be used to provide an objective basis for evaluating a wide range of tree counting and detection processes in future research.

Long-Term Changes in Fusiform Rust Incidence in the Southeastern United States

Fusiform rust is the most devastating disease of slash pine (Pinus elliottii) and loblolly pine (Pinus taeda )i n the southeastern United States. Since the 1970s, the USDA Forest Service Forest Inventory and Analysis (FIA) Program has assessed fusiform rust incidence on its network of ground plots in 13 states across the southeastern United States. Through analysis of the FIA data, we found that current fusiform rust incidence varied by state, forest type, and stand origin and that across all stand ages, rust incidence was approximately equal in planted and natural stands of loblolly pine but was higher for planted versus natural stands of slash pine. Decreases in rust incidence over the last 30 – 40 years were evident in young planted loblolly pine stands but not in young planted slash pine stands. Results for slash pine were surprising, and the reasons remain unclear but one reason may be planting stock origin, which was unknown and may be highly variable in rust resistance. These analyses of FIA rust incidence data also were used to update the original rust disease hazard maps published by Starkey et al. (1997).

Combining satellite lidar, airborne lidar, and ground plots to estimate the amount and distribution of aboveground biomass in the boreal forest of North America

We report estimates of the amount, distribution, and uncertainty of aboveground biomass (AGB) of the different ecoregions and forest land cover classes within the North American boreal forest, analyze the factors driving the error estimates, and compare our estimates with other reported values. A three-phase sampling strategy was used (i) to tie ground plot AGB to airborne profiling lidar metrics and (ii) to link the airborne estimates of AGB to ICESat-GLAS lidar measurements such that (iii) GLAS could be used as a regional sampling tool. We estimated the AGB of the North American boreal forest at 21.8 Pg, with relative error of 1.9% based on 256 GLAS orbits (229 086 pulses). The distribution of AGB was 46.6% for western Canada, 43.7% for eastern Canada, and 9.7% for Alaska. With a single exception, relative errors were under 4% for the three regions and for the major cover types and under 10% at the ecoregion level. The uncertainties of the estimates were calculated using a variance estimator that accounted for only sampling error, i.e., the variability among GLAS orbital estimates, and airborne to spaceborne regression error, i.e., the uncertainty of the model coefficients. Work is ongoing to develop robust statistical techniques for integrating other sources of error such as ground to air regression error and allometric error. Small ecoregions with limited east–west extents tended to have fewer GLAS orbits and a greater percent sampling error. AGB densities derived from GLAS agreed closely with the estimates derived from both forest inventories (<17%) and a MODIS-based interpolation technique (<26%) for more southern, well-inventoried ecoregions, whereas differences were much greater for unmanaged northern and (or) mountainous ecoregions.

Adapting forest management to climate change using bioclimate models with topographic drivers

Bioclimate models incorporating topographic predictors as surrogates for microclimate effects are developed for Populus tremuloides and Picea engelmannii to provide the fine-grained specificity to local terrain required for adapting management of three Colorado (USA) national forests (1.28 million ha) and their periphery to climate change. Models were built with the Random Forests classification tree using presence–absence observations obtained by overlaying species distribution maps on data points gridded at ∼225 m within the forests and from ground plot observations from adjacent areas. Topographic effects derived from 90-m elevation grids were expressed by weighting aspect by slope angle. Climate estimates were obtained from spline surfaces. Out-of-bag errors were ∼17 per cent, and classification errors for an independent sample from within the forest were ∼13 per cent. Topographic variables were second in importance to climate variables for predicting species distributions; their inclusion captured well-known topographic effects on vegetation in mountainous terrain. Predictions made for future climates described by three General Circulation Models and three emissions scenarios were used to map on 90-m grids the habitat expected to be lost, threatened, persistent or emergent. The habitat categories are used to identify those areas where treatments should have highest likelihood of success.

Effects of field plot size on prediction accuracy of aboveground biomass in airborne laser scanning-assisted inventories in tropical rain forests of Tanzania.

BackgroundAirborne laser scanning (ALS) has recently emerged as a promising tool to acquire auxiliary information for improving aboveground biomass (AGB) estimation in sample-based forest inventories. Under design-based and model-assisted inferential frameworks, the estimation relies on a model that relates the auxiliary ALS metrics to AGB estimated on ground plots. The size of the field plots has been identified as one source of model uncertainty because of the so-called boundary effects which increases with decreasing plot size. Recent research in tropical forests has aimed to quantify the boundary effects on model prediction accuracy, but evidence of the consequences for the final AGB estimates is lacking. In this study we analyzed the effect of field plot size on model prediction accuracy and its implication when used in a model-assisted inferential framework.ResultsThe results showed that the prediction accuracy of the model improved as the plot size increased. The adjusted R2 increased from 0.35 to 0.74 while the relative root mean square error decreased from 63.6 to 29.2%. Indicators of boundary effects were identified and confirmed to have significant effects on the model residuals. Variance estimates of model-assisted mean AGB relative to corresponding variance estimates of pure field-based AGB, decreased with increasing plot size in the range from 200 to 3000 m2. The variance ratio of field-based estimates relative to model-assisted variance ranged from 1.7 to 7.7.ConclusionsThis study showed that the relative improvement in precision of AGB estimation when increasing field-plot size, was greater for an ALS-assisted inventory compared to that of a pure field-based inventory.

Detecting the Phenology and Discriminating Mediterranean Natural Habitats With Multispectral Sensors—An Analysis Based on Multiseasonal Field Spectra

Due to their high degree of vegetation heterogeneity, fragmentation, and biodiversity, Mediterranean natural habitats are difficult to assess and monitor with in situ observations solely. Together with standardized ground plots and regular in situ measurements, remote sensing contributes to better understand the diversity of these habitats and their phenology. We used field spectroradiometry to simulate the radiometric signal corresponding to six multispectral satellites: 1) IKONOS, 2) Landsat 5 TM, 3) Landsat 8, 4) Pléiades, 5) Sentinel-2, and 6) WorldView-2. We compared the suitability of each sensor for the estimation of the cover fraction of photosynthetic vegetation (PV) observed for five types of habitats during a vegetation cycle from February to October 2013. We also analyzed the contribution of multiseasonal satellite acquisitions for habitat discrimination. We showed that multivariate regression applied to Worldview-2 reflectance produces the most accurate PV. This was explained by the higher number of spectral bands in the visible domain. Habitat discrimination based on monotemporal acquisitions showed better performances when PV was higher. Sentinel-2 and WorldView-2 outperformed other sensors for each individual date. Multitemporal acquisitions outperformed monotemporal acquisition for habitat discrimination. However, selecting all reflectance data acquired during the season resulted in suboptimal performances compared to more parsimonious combinations. Finally, all of them ranged between 86.6% and 89.2% classification accuracy with multiseasonal acquisitions. New strategies need to be designed to identify individual habitats of particular interest. Defining optimal multiseasonal remote-sensing acquisitions specific to each habitat and appropriate spectral and spatial resolution will contribute to improved discrimination of Mediterranean natural habitats.

Performance of the Enhanced Vegetation Index to Detect Inner-annual Dry Season and Drought Impacts on Amazon Forest Canopies

Abstract. The Amazon rainforests represent the largest connected forested area in the tropics and play an integral role in the global carbon cycle. In the last years the discussion about their phenology and response to drought has intensified. A recent study argued that seasonality in greenness expressed as Enhanced Vegetation Index (EVI) is an artifact of variations in sun-sensor geometry throughout the year. We aimed to reproduce these results with the Moderate-Resolution Imaging Spectroradiometer (MODIS) MCD43 product suite, which allows modeling the Bidirectional Reflectance Distribution Function (BRDF) and keeping sun-sensor geometry constant. The derived BRDF-adjusted EVI was spatially aggregated over large areas of central Amazon forests. The resulting time series of EVI spanning the 2000-2013 period contained distinct seasonal patterns with peak values at the onset of the dry season, but also followed the same pattern of sun geometry expressed as Solar Zenith Angle (SZA). Additionally, we assessed EVI’s sensitivity to precipitation anomalies. For that we compared BRDF-adjusted EVI dry season anomalies to two drought indices (Maximum Cumulative Water Deficit, Standardized Precipitation Index). This analysis covered the whole of Amazonia and data from the years 2000 to 2013. The results showed no meaningful connection between EVI anomalies and drought. This is in contrast to other studies that investigate the drought impact on EVI and forest photosynthetic capacity. The results from both sub-analyses question the predictive power of EVI for large scale assessments of forest ecosystem functioning in Amazonia. Based on the presented results, we recommend a careful evaluation of the EVI for applications in tropical forests, including rigorous validation supported by ground plots.

The importance of tree species and soil taxonomy to modeling forest soil carbon stocks in Canada

Accurate initialization of soil and dead organic matter carbon (C) stocks in forest ecosystem models is challenging but critical to forest C estimation, assessing current and future responses to climate change, and evaluation of management options for climate change mitigation strategies. We identified opportunities to improve the accuracy of soil C estimates from the Carbon Budget of the Canadian Forest Sector (CBM-CFS3) — a model of forest C dynamics used to support greenhouse gas emission reporting. Accuracy of soil C stocks estimated by models is very dependent on the initialization process. Here, we used redundancy analysis (RDA) and ordinations in an exploratory analysis to compare the variance structures of soil C estimates determined by model variables used in the initialization process, in two different soil C datasets; one derived from the model, the other obtained from 2391 ground plots. We also used the ground plot data to determine if soil taxonomy (information currently not used in the CBM-CFS3) could be used to explain variation in addition to that already accounted for by variables in the model. Total variance of the plot C dataset was about twice as large as the variance of the model C dataset confirming that currently the model does not represent all factors that control variation in soil C stocks. Soil C stocks in the mineral soil were highly correlated with C stocks in soil organic horizons in the model dataset but not in the plot dataset, suggesting that the variables included in our assessment controlling C stocks in the mineral soil horizons are different than in the organic soil horizons. Tree productivity (maximum yield curve volume per hectare) explained a much larger proportion of the total variation in the model dataset than in the plot dataset, whereas the leading tree species explained more variation in the plot dataset than in the model, suggesting that accuracy of initialization of soil C stocks could be improved by including leading tree species to stratify soil C modeling parameters. Leading species that are in greatest need of improved representation were identified by ordination. The results from the RDA showed that soil taxonomy explained 4 (order) to 13% (subgroup) of plot soil C variance, in addition to that explained by variables currently used in the model that determine initial soil C stocks. Soil taxonomy and leading species can compensate for one another to explain variance in soil C stocks. Our results suggest the potential of using the combination of leading tree species and soil taxonomy to improve soil C stocks initialized by forest C models, but this remains to be tested.

Automate 3D Modeling of Trace and Via Structures

The simulation and modeling of advanced packaging has always been hindered by the duality of the design process. Electrical design packages are based on vectors, plotting, and routing compounds that align well with the physical process of fabricating substrates. However, these design files contain numerous flaws and limitations that does not allow them to be directly imported into modern CAE tools for mechanical, thermal, and electrical modeling and simulation. The resulting gap results in extended model development (physical artifacts are recreated in the CAE software), gross assumptions about composite properties, or a preference for physical testing. This limitation does not align with the strategic goals of the semiconductor packaging industry, especially in regards to the system-in-package and package-on-package technologies. This paper will discuss a new parsing methodology developed to extract trace, via, and ground plots from standard electronic CAD output formats (Gerber, ODB++, ODB.xml). The methodology for capturing trace geometry and eliminating design flaws and singularities will be explained. The process includes selection of appropriate elements for trace, via, power plane, epoxy, and solder mask to ensure successful automesh completion and segmented hole creation and checking to ensure compatibility with model development strategies implemented in Abaqus and Ansys software packages. Identification of material properties through rule checks and computation of effective properties using sandwich and Halpern-Tsai equations will be demonstrated. The presentation will conclude with a discussion on case studies where automated 3D modeling of all copper features provides strong advantages over existing processes, including warpage prediction of wafers, strips, and packages, effective CTE calculations, and high frequency electrical signal simulations.