Olive tree growing is an important economic activity in many countries, mostly in the Mediterranean Basin, Argentina, Chile, Australia, and California. Although recent intensification techniques ...organize olive groves in hedgerows, most olive groves are rainfed and the trees are scattered (as in Spain and Italy, which account for 50% of the world's olive oil production). Accurate measurement of trees biovolume is a first step to monitor their performance in olive production and health. In this work, we use one of the most accurate deep learning instance segmentation methods (Mask R-CNN) and unmanned aerial vehicles (UAV) images for olive tree crown and shadow segmentation (OTCS) to further estimate the biovolume of individual trees. We evaluated our approach on images with different spectral bands (red, green, blue, and near infrared) and vegetation indices (normalized difference vegetation index-NDVI-and green normalized difference vegetation index-GNDVI). The performance of red-green-blue (RGB) images were assessed at two spatial resolutions 3 cm/pixel and 13 cm/pixel, while NDVI and GNDV images were only at 13 cm/pixel. All trained Mask R-CNN-based models showed high performance in the tree crown segmentation, particularly when using the fusion of all dataset in GNDVI and NDVI (F1-measure from 95% to 98%). The comparison in a subset of trees of our estimated biovolume with ground truth measurements showed an average accuracy of 82%. Our results support the use of NDVI and GNDVI spectral indices for the accurate estimation of the biovolume of scattered trees, such as olive trees, in UAV images.
There is a growing demand for accurate high-resolution land cover maps in many fields, e.g., in land-use planning and biodiversity conservation. Developing such maps has been traditionally performed ...using Object-Based Image Analysis (OBIA) methods, which usually reach good accuracies, but require a high human supervision and the best configuration for one image often cannot be extrapolated to a different image. Recently, deep learning Convolutional Neural Networks (CNNs) have shown outstanding results in object recognition in computer vision and are offering promising results in land cover mapping. This paper analyzes the potential of CNN-based methods for detection of plant species of conservation concern using free high-resolution Google Earth TM images and provides an objective comparison with the state-of-the-art OBIA-methods. We consider as case study the detection of Ziziphus lotus shrubs, which are protected as a priority habitat under the European Union Habitats Directive. Compared to the best performing OBIA-method, the best CNN-detector achieved up to 12% better precision, up to 30% better recall and up to 20% better balance between precision and recall. Besides, the knowledge that CNNs acquired in the first image can be re-utilized in other regions, which makes the detection process very fast. A natural conclusion of this work is that including CNN-models as classifiers, e.g., ResNet-classifier, could further improve OBIA methods. The provided methodology can be systematically reproduced for other species detection using our codes available through (https://github.com/EGuirado/CNN-remotesensing).
Vegetation generally appears scattered in drylands. Its structure, composition and spatial patterns are key controls of biotic interactions, water, and nutrient cycles. Applying segmentation methods ...to very high-resolution images for monitoring changes in vegetation cover can provide relevant information for dryland conservation ecology. For this reason, improving segmentation methods and understanding the effect of spatial resolution on segmentation results is key to improve dryland vegetation monitoring. We explored and analyzed the accuracy of Object-Based Image Analysis (OBIA) and Mask Region-based Convolutional Neural Networks (Mask R-CNN) and the fusion of both methods in the segmentation of scattered vegetation in a dryland ecosystem. As a case study, we mapped
, the dominant shrub of a habitat of conservation priority in one of the driest areas of Europe. Our results show for the first time that the fusion of the results from OBIA and Mask R-CNN increases the accuracy of the segmentation of scattered shrubs up to 25% compared to both methods separately. Hence, by fusing OBIA and Mask R-CNNs on very high-resolution images, the improved segmentation accuracy of vegetation mapping would lead to more precise and sensitive monitoring of changes in biodiversity and ecosystem services in drylands.
Abstract
Land-Use and Land-Cover (LULC) mapping is relevant for many applications, from Earth system and climate modelling to territorial and urban planning. Global LULC products are continuously ...developing as remote sensing data and methods grow. However, there still exists low consistency among LULC products due to low accuracy in some regions and LULC types. Here, we introduce Sentinel2GlobalLULC, a Sentinel-2 RGB image dataset, built from the spatial-temporal consensus of up to 15 global LULC maps available in Google Earth Engine. Sentinel2GlobalLULC v2.1 contains 194877 single-class RGB image tiles organized into 29 LULC classes. Each image is a 224 × 224 pixels tile at 10 × 10 m resolution built as a cloud-free composite from Sentinel-2 images acquired between June 2015 and October 2020. Metadata includes a unique LULC annotation per image, together with level of consensus, reverse geo-referencing, global human modification index, and number of dates used in the composite. Sentinel2GlobalLULC is designed for training deep learning models aiming to build precise and robust global or regional LULC maps.
Climate change has been shown to alter the spatial distribution of whales and other marine mammals. Fast changing ocean temperatures may also affect the spatial distribution of whales at a finer ...scale, namely within populations, including aggregation behaviour. Our ability to analyze the impact of climate change on whale aggregation behavior, however, has been limited by our ability to collect spatial observation data over large areas. To overcome this limitation, this study analyzed open-access satellite imagery obtained between 2007 and 2020 in Canada, Russia, and Alaska using Deep Convolutional Neural Networks (CNN) to detect 1,980 beluga whales in 11 populations and to quantify their aggregation patterns within their populations. Subsequently, we examined the relationship between sea surface temperature (SST) and the intra-population spatial patterns of beluga whales during summer seasons, when these whales normally aggregate. We detected a negative correlation between SST and the frequency of beluga whale aggregation, suggesting that warming temperatures may impact beluga whale spatio-behavioral dynamics. Considering that the relative abundance of beluga whales is declining and the future SST projections in these Arctic Ocean locations, climate change may pose yet another threat to beluga whales and other ice-dependent species.
Accurate tree cover mapping is of paramount importance in many fields, from biodiversity conservation to carbon stock estimation, ecohydrology, erosion control, or Earth system modelling. Despite ...this importance, there is still uncertainty about global forest cover, particularly in drylands. Recently, the Food and Agriculture Organization of the United Nations (FAO) conducted a costly global assessment of dryland forest cover through the visual interpretation of orthoimages using the Collect Earth software, involving hundreds of operators from around the world. Our study proposes a new automatic method for estimating tree cover using artificial intelligence and free orthoimages. Our results show that our tree cover classification model, based on convolutional neural networks (CNN), is 23% more accurate than the manual visual interpretation used by FAO, reaching up to 79% overall accuracy. The smallest differences between the two methods occurred in the driest regions, but disagreement increased with the percentage of tree cover. The application of CNNs could be used to improve and reduce the cost of tree cover maps from the local to the global scale, with broad implications for research and management.
Biocontrol bacteria can be used for plant protection against some plant diseases. Pseudomonas chlororaphis PCL1606 (PcPCL1606) is a model bacterium isolated from the avocado rhizosphere with strong ...antifungal antagonism mediated by the production of 2-hexyl, 5-propil resorcinol (HPR). Additionally, PcPCL1606 has biological control against different soil-borne fungal pathogens, including the causal agent of the white root rot of many woody crops and avocado in the Mediterranean area, Rosellinia necatrix. The objective of this study was to assess whether the semicommercial application of PcPCL1606 to soil can potentially affect avocado soil and rhizosphere microbial communities and their activities in natural conditions and under R. necatrix infection. To test the putative effects of PcPCL1606 on soil eukaryotic and prokaryotic communities, a formulated PcPCL1606 was prepared and applied to the soil of avocado plants growing in mesocosm experiments, and the communities were analyzed by using 16S/ITS metagenomics. PcPCL1606 survived until the end of the experiments. The effect of PcPCL1606 application on prokaryotic communities in soil and rhizosphere samples from natural soil was not detectable, and very minor changes were observed in eukaryotic communities. In the infested soils, the presence of R. necatrix strongly impacted the soil and rhizosphere microbial communities. However, after PcPCL1606 was applied to soil infested with R. necatrix, the prokaryotic community reacted by increasing the relative abundance of few families with protective features against fungal soilborne pathogens and organic matter decomposition (Chitinophagaceae, Cytophagaceae), but no new prokaryotic families were detected. The treatment of PcPCL1606 impacted the fungal profile, which strongly reduced the presence of R. necatrix in avocado soil and rhizosphere, minimizing its effect on the rest of the microbial communities. The bacterial treatment of formulated PcPCL1606 on avocado soils infested with R. necatrix resulted in biological control of the pathogen. This suppressiveness phenotype was analyzed, and PcPCL1606 has a key role in suppressiveness induction; in addition, this phenotype was strongly dependent on the production of HPR.
From its origins, the concept of desertification has been controversial. The prevailing confusion between two desertification visions, one that considers it as the expansion of deserts and another ...that emphasizes its anthropogenic component, has been transferred to society. Here we illustrate misunderstandings about desertification using a very illustrative case from the Tabernas-Sorbas Basin (Almeria, Spain), where striking badlands that are often used as an image of desertification coexist with an intensive olive agriculture that is irreversibly deteriorating the only oasis in continental Europe (Los Molinos spring). The olive tree is a traditional Mediterranean dryland crop and until the 1950s only about 200 ha were irrigated in this area. However, the profitability of the crop has caused irrigation to expand to 4400 ha in the last two decades. The process of intensification has been reinforced giving way to super-intensive irrigation, which involves going from 210 to 1550 trees/ha, which in a few years already occupies more than 1500 ha. The effects on the water balance of the aquifer feeding these crops have been severe, and the flow of the Los Molinos spring has gone from more than 40 L/s for the period 1970–2000 to the current 7.28 L/s. Unraveling the mechanisms of land degradation and its main drivers are the first step to propose management actions to achieve a more sustainable use of resources and to combat desertification.
The stabling of livestock farming implies changes in both local ecosystems (regeneration of forest stands via reduced grazing) and those located thousands of kilometers away (deforestation to produce ...grain for feeding livestock). Despite their importance, these externalities are poorly known. Here we evaluated how the intensification and confinement of livestock in Spain has affected forest surface changes there and in South America, the largest provider of soybeans for animal feed to the European Union. For this purpose, we have used Spanish soybean import data from Brazil, Paraguay and Argentina and a land condition map of Spain. The area of secondary forest in Spain that has regenerated as a result of livestock stabling has been ~7000 kha for the decade 2000–2010. In the same period, 1220 kha of high value South American ecosystems (e.g., Chaco dry Forest, Amazonian rainforest or Cerrado) have been deforested. While these figures may offer a favorable interpretation of the current industrial livestock production, it is not possible to speak of compensation when comparing the destruction of well-structured ecosystems, such as primary South American forests, with the creation of secondary forest landscapes in Spain, which are also prone to wildfires. Our results highlight how evaluating land use change policies at a national or regional level is an incomplete exercise in our highly telecoupled and globalized world.
The latest world atlas of desertification represents a turning point in the diagnosis of desertification. While it forgoes desertification mapping due to the intrinsic complexity of the phenomenon ...and the impossibility of measuring it using a single indicator, it introduces the convergence of evidence paradigm, which identifies socioeconomic and biophysical variables whose behaviour allows pointing out those areas prone to desertification. The Spanish National Action Program Against Desertification (PAND), back in 2008, already implemented a similar approach to identify five “desertification landscapes” within Spain using both socio-economic and climatic information. The PAND was not only pioneering but also, unfortunately, accurate. Desertification in Spain has continued to worsen and the first two decades of the 21st century have consolidated an agri-food model whose dynamics have exacerbated the desertification processes identified in the PAND. Despite its scientific value, the PAND lacked a proper action plan and was completely detached from the diagnosis. As a result, the diagnosis it provided was not followed by effective actions to halt desertification in Spain. The Spanish government’s recent declaration of climate and environmental emergency requires a new strategy to combat desertification. This commitment is an excellent opportunity to update the diagnosis of the situation and, more crucially, to unify the different desertification sectoral policies and actions under a single front. We provide here elements (e.g., analysis of agri-food trends and integration of plans and policies at different geographical and sectoral levels) for a roadmap to be designed around the pressures, impacts, and drivers that define today’s Spanish desertification landscapes to effectively manage and avoid their further degradation.