Ozone impact on Mediterranean forests remains largely under-investigated, despite strong photochemical activity and harmful effects on crops. As representative of O
3 impacts on Mediterranean ...vegetation, this paper reviews the current knowledge about O
3 and forests in Italy. The intermediate position between Africa and European mid-latitudes creates a complex patchwork of climate and vegetation. Available data from air quality monitoring stations and passive samplers suggest O
3 levels regularly exceed the critical level (CL) for forests. In contrast, relationships between O
3 exposure and effects (crown transparency, radial growth and foliar visible symptoms) often fail. Despite limitations in the study design or underestimation of the CL can also affect this discrepancy, the effects of site factors and plant ecology suggest Mediterranean forest vegetation is adapted to face oxidative stress, including O
3. Implications for risk assessment (flux-based CL, level III, non-stomatal deposition) are discussed.
Why Mediterranean forests are more ozone tolerant than mesophilic vegetation is explored.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK
Convolutional neural networks (CNNs) exhibit good performance in image processing tasks, pointing themselves as the current state-of-the-art of deep learning methods. However, the intrinsic ...complexity of remotely sensed hyperspectral images still limits the performance of many CNN models. The high dimensionality of the HSI data, together with the underlying redundancy and noise, often makes the standard CNN approaches unable to generalize discriminative spectral-spatial features. Moreover, deeper CNN architectures also find challenges when additional layers are added, which hampers the network convergence and produces low classification accuracies. In order to mitigate these issues, this paper presents a new deep CNN architecture specially designed for the HSI data. Our new model pursues to improve the spectral-spatial features uncovered by the convolutional filters of the network. Specifically, the proposed residual-based approach gradually increases the feature map dimension at all convolutional layers, grouped in pyramidal bottleneck residual blocks, in order to involve more locations as the network depth increases while balancing the workload among all units, preserving the time complexity per layer. It can be seen as a pyramid, where the deeper the blocks, the more feature maps can be extracted. Therefore, the diversity of high-level spectral-spatial attributes can be gradually increased across layers to enhance the performance of the proposed network with the HSI data. Our experiments, conducted using four well-known HSI data sets and 10 different classification techniques, reveal that our newly developed HSI pyramidal residual model is able to provide competitive advantages (in terms of both classification accuracy and computational time) over the state-of-the-art HSI classification methods
Capsule Networks for Hyperspectral Image Classification Paoletti, Mercedes E.; Haut, Juan Mario; Fernandez-Beltran, Ruben ...
IEEE transactions on geoscience and remote sensing,
04/2019, Volume:
57, Issue:
4
Journal Article
Peer reviewed
Open access
Convolutional neural networks (CNNs) have recently exhibited an excellent performance in hyperspectral image classification tasks. However, the straightforward CNN-based network architecture still ...finds obstacles when effectively exploiting the relationships between hyperspectral imaging (HSI) features in the spectral-spatial domain, which is a key factor to deal with the high level of complexity present in remotely sensed HSI data. Despite the fact that deeper architectures try to mitigate these limitations, they also find challenges with the convergence of the network parameters, which eventually limit the classification performance under highly demanding scenarios. In this paper, we propose a new CNN architecture based on spectral-spatial capsule networks in order to achieve a highly accurate classification of HSIs while significantly reducing the network design complexity. Specifically, based on Hinton's capsule networks, we develop a CNN model extension that redefines the concept of capsule units to become spectral-spatial units specialized in classifying remotely sensed HSI data. The proposed model is composed by several building blocks, called spectral-spatial capsules, which are able to learn HSI spectral-spatial features considering their corresponding spatial positions in the scene, their associated spectral signatures, and also their possible transformations. Our experiments, conducted using five well-known HSI data sets and several state-of-the-art classification methods, reveal that our HSI classification approach based on spectral-spatial capsules is able to provide competitive advantages in terms of both classification accuracy and computational time.
Super-resolution (SR) brings an excellent opportunity to improve a wide range of different remote sensing applications. SR techniques are concerned about increasing the image resolution while ...providing finer spatial details than those captured by the original acquisition instrument. Therefore, SR techniques are particularly useful to cope with the increasing demand remote sensing imaging applications requiring fine spatial resolution. Even though different machine learning paradigms have been successfully applied in SR, more research is required to improve the SR process without the need of external high-resolution (HR) training examples. This paper proposes a new convolutional generator model to super-resolve low-resolution (LR) remote sensing data from an unsupervised perspective. That is, the proposed generative network is able to initially learn relationships between the LR and HR domains throughout several convolutional, downsampling, batch normalization, and activation layers. Then, the data are symmetrically projected to the target resolution while guaranteeing a reconstruction constraint over the LR input image. An experimental comparison is conducted using 12 different unsupervised SR methods over different test images. Our experiments reveal the potential of the proposed approach to improve the resolution of remote sensing imagery.
Convolutional neural networks (CNNs) have become quite popular for solving many different tasks in remote sensing data processing. The convolution is a linear operation, which extracts features from ...the input data. However, nonlinear operations are able to better characterize the internal relationships and hidden patterns within complex remote sensing data, such as hyperspectral images (HSIs). Morphological operations are powerful nonlinear transformations for feature extraction that preserve the essential characteristics of the image, such as borders, shape, and structural information. In this article, a new end-to-end morphological deep learning framework (called MorphConvHyperNet ) is introduced. The proposed approach efficiently models nonlinear information during the training process of HSI classification. Specifically, our method includes spectral and spatial morphological blocks to extract relevant features from the HSI input data. These morphological blocks consist of two basic 2-D morphological operators (erosion and dilation) in the respective layers, followed by a weighted combination of the feature maps. Both layers can successfully encode the nonlinear information related to shape and size, playing an important role in classification performance. Our experimental results, obtained on five widely used HSIs, reveal that our newly proposed MorphConvHyperNet offers comparable (and even superior) performance when compared to traditional 2-D and 3-D CNNs for HSI classification.
•Water vapour and ozone fluxes were measured above and below a canopy of Holm Oak.•We modelled ozone deposition to soil, stomata, and cuticles.•Stomata explained almost the totality of ozone fluxes ...during the winter.•Soils removed up to 30% of ozone.
Castelporziano Estate is a coastal forest 25km from downtown Rome. It is an ideal site to study interactions between Mediterranean forest ecosystems and a polluted atmosphere. Two eddy covariance systems were used to simultaneously measure water vapour and ozone fluxes above and below a canopy of Holm Oak (Quercus ilex). Additional measurements of environmental parameters allowed to calculate stomatal ozone fluxes in order to parameterize atmospheric models and new algorithms for discriminating ozone deposition into its three more significant sinks: soil, stomata, and cuticles. Results showed that stomata explained almost the totality of ozone fluxes during the winter season, and <60% during the warm seasons under condition of drought stress. Soils removed up to 30% of ozone, suggesting the importance of this sink in this forest ecosystem. This study spanning all seasons over a 2-year period advanced our understanding about the contribution of a representative Mediterranean Oak forest to biosphere–atmosphere exchange.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
European standards for the protection of forests from ozone (O3) are based on atmospheric exposure (AOT40) that is not always representative of O3 effects since it is not a proxy of gas uptake ...through stomata (stomatal flux). MOTTLES “MOnitoring ozone injury for seTTing new critical LEvelS” is a LIFE project aimed at establishing a permanent network of forest sites based on active O3 monitoring at remote areas at high and medium risk of O3 injury, in order to define new standards based on stomatal flux, i.e. PODY (Phytotoxic Ozone Dose above a threshold Y of uptake). Based on the first year of data collected at MOTTLES sites, we describe the MOTTLES monitoring station, together with protocols and metric calculation methods. AOT40 and PODY, computed with different methods, are then compared and correlated with forest–health indicators (radial growth, crown defoliation, visible foliar O3 injury). For the year 2017, the average AOT40 calculated according to the European Directive was even 5 times (on average 1.7 times) the European legislative standard for the protection of forests. When the metrics were calculated according to the European protocols (EU Directive 2008/50/EC or Modelling and Mapping Manual LTRAP Convention), the values were well correlated to those obtained on the basis of the real duration of the growing season (i.e. MOTTLES method) and were thus representative of the actual exposure/flux. AOT40 showed opposite direction relative to PODY. Visible foliar O3 injury appeared as the best forest–health indicator for O3 under field conditions and was more frequently detected at forest edge than inside the forest. The present work may help the set–up of further long–term forest monitoring sites dedicated to O3 assessment in forests, especially because flux-based assessments are recommended as part of monitoring air pollution impacts on ecosystems in the revised EU National Emissions Ceilings Directive.
•The MOTTLES network for active O3 monitoring in forests is described.•In 2017, AOT40 exceeded twice the limit of the European Directive for forests.•O3 metrics from European protocols were representative of actual exposure/fluxes.•AOT40 and PODy were inversely correlated.•Visible foliar injury was the best forest–health indicator for O3.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Nowadays, a large number of remote sensing instruments are providing a massive amount of data within the frame of different Earth Observation missions. These instruments are characterized by the wide ...variety of data they can collect, as well as the impressive volume of data and the speed at which it is acquired. In this sense, hyperspectral imaging data has certain properties that make it difficult to process, such as its large spectral dimension coupled with problematic data variability. To overcome these challenges, convolutional neural networks have been proposed as classification models because of their ability to extract relevant spectral–spatial features and learn hidden patterns, along their great architectural flexibility. Their high performance relies on the convolution kernels to exploit the spatial relationships. Thus, filter design is crucial for the correct performance of models. Nevertheless, hyperspectral data may contain objects with different shapes and orientations, preventing filters from “seeing everything possible” during the decision making. To overcome this limitation, this paper proposes a novel adaptable convolution model based on deforming kernels combined with deforming convolution layers to fit their effective receptive field to the input data. The proposed adaptable convolutional network (named DKDCNet) has been evaluated over two well-known hyperspectral scenes, demonstrating that it is able to achieve better results than traditional strategies with similar computational cost for HSI classification.
Full text
Available for:
IZUM, KILJ, NUK, PILJ, PNG, SAZU, UL, UM, UPUK
Emission of BVOC (Biogenic Volatile Organic Compounds) from plant leaves in response to ozone exposure (O3) and nitrogen (N) fertilization is poorly understood. For the first time, BVOC emissions ...were explored in a forest tree species (silver birch, Betula pendula) exposed for two years to realistic levels of O3 (35, 48 and 69 ppb as daylight average) and N (10, 30 and 70 kg ha−1 yr−1, applied weekly to the soil as ammonium nitrate). The main BVOCs emitted were: α-pinene, β-pinene, limonene, ocimene, (E)-4,8-dimethyl-1,3,7-nonatriene (DMNT) and hexanal. Ozone exposure increased BVOC emission and reduced total leaf area. The effect on emission was stronger when a short-term O3 metric (concentrations at the time of sampling) rather than a long-term one (AOT40) was used. The effect of O3 on total leaf area was not able to compensate for the stimulation of emission, so that responses to O3 at leaf and whole-plant level were similar. Nitrogen fertilization increased total leaf area, decreased α-pinene and β-pinene emission, and increased ocimene, hexanal and DMNT emission. The increase of leaf area changed the significance of the emission response to N fertilization for most compounds. Nitrogen fertilization mitigated the effects of O3 exposure on total leaf area, while the combined effects of O3 exposure and N fertilization on BVOC emission were additive and not synergistic. In conclusion, O3 exposure and N fertilization have the potential to affect global BVOC via direct effects on plant emission rates and changes in leaf area.
Display omitted
•Ozone exposure stimulated BVOC emission.•Responses to an episodic O3 metric were stronger than to a seasonal one.•N fertilization showed compound-specific effects (from inhibition to stimulation).•N-driven changes in total leaf area affected the magnitude of these effects.•O3 and N impacted BVOC emission individually, with no significant interactions.
Ozone exposure increased BVOC emission and reduced total leaf area, while N fertilization increased total leaf area and showed compound-specific effects (from inhibition to stimulation) on BVOC emission.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK
Present standards for protecting ecosystems from ozone (O3), such as AOT40, use atmospheric concentrations. A stomatal flux-based approach (Phytotoxic O3 Dose, PODY) has been suggested. We compared ...the spatial and temporal distribution of AOT40 and PODY – with and without a hourly threshold of uptake (POD1 and POD0) – for Pinus halepensis and Fagus sylvatica in South-eastern France and North-western Italy. Ozone uptake was simulated by including limitation due to soil water content, as this is an important parameter in water-limited environments. Both AOT40 and POD1 exceeded the critical levels suggested for forests. AOT40 suggested a larger O3 risk relative to PODY. No significant spatial and temporal difference occurred between POD1 and POD0. The use of POD0 in the assessment of ambient O3 risk for vegetation is thus recommended, because it is more biologically-meaningful than AOT40 and easier to be calculated than POD1. Canopy Moisture Content (CMC), a proxy of foliar water content, was modelled and tested as a potential plant O3 response indicator. CMC response to O3 was species-specific, and thus cannot be recommended in the epidemiology of O3 injury to forests.
Display omitted
•Stomatal ozone flux (POD) is a better metric than AOT40 for the protection of forests.•We recommend POD0 rather than POD1.•We cannot recommend CMC as a plant-response indicator.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UL, UM, UPCLJ, UPUK
PDF
