Mapping the pyrogenic dynamics of forest geosystems on the northeastern shore of Lake Baikal

The paper considers the impact of the pyrogenic factor on the landscapes of the Barguzinskii Range. Model representative sites with natural and disturbed mountain-taiga geosystems are identified in the territory of the Trans-Baikal National Park (Svyatoi Nos Peninsula) and the Barguzinskii Nature Reserve (Shumilikha, Tarkulik, and Davsha river valleys). We used geoinformation methods, landscape interpretation mapping, field observation data, remote sensing data and traditional comparative geographical methods for assessment and mapping. The collected data are systematized in the form of a geoinformation database for individual sections and visualized in a cartographic form. We compiled vegetation maps, taking into account the features of the relief and soil types and gave a general description of the landscape state to analyze the pyrogenic impact on local landscapes. It was also revealed that the modification processes of forest geosystems caused by the pyrogenic impact are widespread across wide swathes, but have a different character. It depends on the individual spatial geographical features of the selected representative sites and the nature and time of the direct pyrogenic factor exposure, e.g., the frequency, intensity, especially the microclimate, relief, etc. The paper shows that the restorative stages of plant dynamics in the model sites are clearly traced. We have established an insignificant difference between the current and reference states in places of weak pyrogenic impact, significant local state changes in places of extensive areal impact, and significant and catastrophic changes in places of lasting and intense pyrogenic impact. Maps of the geosystem disturbance caused by both pyrogenic and natural factors for the model sites were compiled.

Mapping vegetation types in Antarctic Peninsula and South Shetlands islands using Sentinel-2 images and Google Earth Engine cloud computing

The Antarctic vegetation maps are usually made using very high-resolution images collected by orbital sensors or unmanned aerial vehicles, generating isolated maps with information valid only for the time of image acquisition. In the context of global environmental change, mapping the current Antarctic vegetation distribution on a regular basis is necessary for a better understanding of the changes in this fragile environment. This work aimed to generate validated vegetation maps for the North Antarctic Peninsula and South Shetlands Islands based on Sentinel-2 images using cloud processing. Sentinel-2 imagery level 1C, acquired between 2016 and 2021 (January-April), were used. Land pixels were masked with the minimum value composite image for the "water vapor" band. The NDVI maximum value composite image was sliced, and its classes were associated with the occurrence of algae (0.15 - 0.20), lichens (0.20 - 0.50), and mosses (0.50 - 0.80). The vegetation map was validated by comparing it with those from the literature. The present study showed that Sentinel-2 images allow building a validated vegetation type distribution map for Antarctica Peninsula and South Shetlands Islands.

Urban Vegetation Mapping from Aerial Imagery Using Explainable AI (XAI)

Urban vegetation mapping is critical in many applications, i.e., preserving biodiversity, maintaining ecological balance, and minimizing the urban heat island effect. It is still challenging to extract accurate vegetation covers from aerial imagery using traditional classification approaches, because urban vegetation categories have complex spatial structures and similar spectral properties. Deep neural networks (DNNs) have shown a significant improvement in remote sensing image classification outcomes during the last few years. These methods are promising in this domain, yet unreliable for various reasons, such as the use of irrelevant descriptor features in the building of the models and lack of quality in the labeled image. Explainable AI (XAI) can help us gain insight into these limits and, as a result, adjust the training dataset and model as needed. Thus, in this work, we explain how an explanation model called Shapley additive explanations (SHAP) can be utilized for interpreting the output of the DNN model that is designed for classifying vegetation covers. We want to not only produce high-quality vegetation maps, but also rank the input parameters and select appropriate features for classification. Therefore, we test our method on vegetation mapping from aerial imagery based on spectral and textural features. Texture features can help overcome the limitations of poor spectral resolution in aerial imagery for vegetation mapping. The model was capable of obtaining an overall accuracy (OA) of 94.44% for vegetation cover mapping. The conclusions derived from SHAP plots demonstrate the high contribution of features, such as Hue, Brightness, GLCM_Dissimilarity, GLCM_Homogeneity, and GLCM_Mean to the output of the proposed model for vegetation mapping. Therefore, the study indicates that existing vegetation mapping strategies based only on spectral characteristics are insufficient to appropriately classify vegetation covers.

A feasibility study on exploitation of atmospheric water resources using cloud seeding in arid and semi arid parts of Iran

Water resource management and optimum use of all available water resources are ways for the adaptation of climate change and drought conditions.Weather modification, commonly known as cloud seeding, is the application of scientific technology that can enhance a cloud's ability to produce precipitation. Cloud seeding projects have being performed in Iran since 1999, as one of the most important strategies to increase water supplies. However, determining the time and place of cloud seeding operation are the basic necessities to achieve the best possible results. This paper explains innovative and scientific methods of feasibility study of Tehran province (semi-arid area) and Hormozgan province (arid area) by meteorological stations data, upper air data, satellite imagery (TRMM), weather radars products, topographic and vegetation maps. Factor analysis and World Meteorological Organization classification methods of PEP were applied to determine proper seeding time. Eventually, cloud seeding susceptible areas were classified as levels and altitudes. In Tehran province, the results showed that the northwest, west and central catchments are capable for cloud seeding operation in December, January, February and March at an altitude of 2500-3000 meters. Suitable months of cloud seeding operation in the north and northwest of the catchment of Hormozgan province were January, February and March, in order of priority.

Assessment of Land-Use Scenarios at a National Scale Using Intensity Analysis and Figure of Merit Components

To address the impacts of future land changes on biodiversity and ecosystem services, land-use scenarios have been developed at the national scale in Japan. However, the validation of land-use scenarios remains a challenge owing to the lack of an appropriate validation method. This research developed land-use maps for 10 land-use categories to calibrate a land-change model for the 1987–1998 period, simulate changes during the 1998–2014 period, and validate the simulation for the 1998–2014 period. Following an established method, this study assessed the three types of land change: (1) reference change during the calibration time interval, (2) simulation change during the validation time interval, and (3) reference change during the validation time interval, using intensity analysis and figure of merit components (hits, misses, and false alarms). The results revealed the cause of the low accuracy of the national scale land-use scenarios as well as priority solutions, such as aligning the underlying spatial vegetation maps and improving the model to reduce two types of disagreement between the simulation and reference maps. These findings should help to improve the accuracy of model predictions and help to better inform policymakers during the decision-making process.

Proximate and Underlying Deforestation Causes in a Tropical Basin through Specialized Consultation and Spatial Logistic Regression Modeling

The present study focuses on identifying and describing the possible proximate and underlying causes of deforestation and its factors using the combination of two techniques: (1) specialized consultation and (2) spatial logistic regression modeling. These techniques were implemented to characterize the deforestation process qualitatively and quantitatively, and then to graphically represent the deforestation process from a temporal and spatial point of view. The study area is the North Pacific Basin, Mexico, from 2002 to 2014. The map difference technique was used to obtain deforestation using the land-use and vegetation maps. A survey was carried out to identify the possible proximate and underlying causes of deforestation, with the aid of 44 specialized government officials, researchers, and people who live in the surrounding deforested areas. The results indicated total deforestation of 3938.77 km2 in the study area. The most important proximate deforestation causes were agricultural expansion (53.42%), infrastructure extension (20.21%), and wood extraction (16.17%), and the most important underlying causes were demographic factors (34.85%), economics factors (29.26%), and policy and institutional factors (22.59%). Based on the spatial logistic regression model, the factors with the highest statistical significance were forestry productivity, the slope, the altitude, the distance from population centers with fewer than 2500 inhabitants, the distance from farming areas, and the distance from natural protected areas.

Correspondence Between Vegetation Patterns and Soils in Wet and Wet-mesic Grasslands of Hanság and Tóköz (Hungary)

Our research focused on the causes responsible for the fine mosaic pattern of plant associations on wet and wet-mesic meadows. The study area is located in the Little Hungarian Plain, including the former swamp basins of Hanság and Tóköz in Hungary. The vegetation survey data were evaluated by statistical methods (TWINSPAN method), and vegetation maps of the areas were prepared. Topsoil samples near the relevés were gathered for further laboratory tests. Soil profiles were opened by a Pürckhauer soil sampler for on-site description of the soil horizons and classification. Surface models provided a base for the preparation of contour maps that could be compared with the vegetation pattern. We found that of the two dominant vegetation types, mesotrophic wet meadows were associated with Mollic Gleysols, while non-tussock sedge beds were associated with Histic Gleysols. At the transitions of the two soil classes, the subgroup of non-tussock sedge beds is the dominant type. The soil class only determined the plant association on a habitat level, but it could not reason the fine pattern of the plant communities on the same soil class. Canonical correspondence analysis (CCA) was performed to investigate the relationship between the distribution of dominant species and soil parameters. Several soil parameters combined have a significant effect on the distribution of dominant species. In conclusion, we found that the formation of association types strongly depends on the soil characteristics of the area, and that it is closely related to it. However, in the formation of the fine mosaic pattern, the driving ecological factors are the microrelief and the length of the saturated or flooded soil conditions.

RATIONALE FOR MONITORING GROUNDWATER LEVELS IN THE REHABILITATED PEATLANDS OF PRIPYAT POLESIE

Based on studies of geomorphology, stratigraphy, hydrology, various maps of Polesie, zones of influence of amelioration canals, vegetation maps, modern satellite images, as well as field studies of peatlands of Pripyat Polesie, two independent drainage systems have been identified, with a network of amelioration canals that intensively discharge water into the rivers Stwiga and Ybort`. Maps of key points were built In GIS-format, on which five sites were laid in the field within the Mezhch and Neresnya peat deposits for further long-term monitoring of GWL parameters. The locations of the sensors installation are based on the relationship of bog phytocenoses with the average annual GWL values and the amplitude of their fluctuations. Analysis of the GWL parameters before and after environmental rehabilitation measures will allow assessing the effectiveness of planned measures in disturbed areas and obtaining new data on areas of peat deposits in their natural state.

Correlation among Earthwork and Cropmark Anomalies within Archaeological Landscape Investigation by Using LiDAR and Multispectral Technologies from UAV

This project aimed to systematically investigate the archaeological remains of the imperial Domitian villa in Sabaudia (Italy), using different three-dimensional survey techniques. Particular attention in the research was paid to the identification and documentation of traces that buried structures left on the surface occupied by the villa, which extended for 46 hectares, an area that was fully covered with structures. Since a dense pine forest was planted during the 1940s and is currently covering the site, this contribution investigates particularly the correlation among the presence of cropmarks, identifiable with the processing of multispectral maps and vegetation indices from RGB images, and earthwork anomalies identified in a Digital Terrain Model (DTM) built, by utilizing a light detection and ranging (LiDAR) flight from an Unmanned Aerial Vehicle (UAV). The study demonstrates how the use of vegetation maps—calculated starting from RGB and multispectral aerial photos—can provide a more expeditious preliminary analysis on the position and extension of areas characterized by the presence of buried structures, but also that, in order to investigate in-depth a context in similar conditions, the most effective approach remains the one based on LiDAR technology. The integration between the two techniques may prove fruitful in limiting the extension of the areas to be investigated with terrestrial survey techniques.

Vegetation classification at Lake Mead National Recreation Area, Mojave National Preserve, Castle Mountains National Monument, and Death Valley National Park: Final report (Revised with Cost Estimate)

Vegetation inventory and mapping is a process to document the composition, distribution and abundance of vegetation types across the landscape. The National Park Service’s (NPS) Inventory and Monitoring (I&M) program has determined vegetation inventory and mapping to be an important resource for parks; it is one of 12 baseline inventories of natural resources to be completed for all 270 national parks within the NPS I&M program. The Mojave Desert Network Inventory & Monitoring (MOJN I&M) began its process of vegetation inventory in 2009 for four park units as follows: Lake Mead National Recreation Area (LAKE), Mojave National Preserve (MOJA), Castle Mountains National Monument (CAMO), and Death Valley National Park (DEVA). Mapping is a multi-step and multi-year process involving skills and interactions of several parties, including NPS, with a field ecology team, a classification team, and a mapping team. This process allows for compiling existing vegetation data, collecting new data to fill in gaps, and analyzing the data to develop a classification that then informs the mapping. The final products of this process include a vegetation classification, ecological descriptions and field keys of the vegetation types, and geospatial vegetation maps based on the classification. In this report, we present the narrative and results of the sampling and classification effort. In three other associated reports (Evens et al. 2020a, 2020b, 2020c) are the ecological descriptions and field keys. The resulting products of the vegetation mapping efforts are, or will be, presented in separate reports: mapping at LAKE was completed in 2016, mapping at MOJA and CAMO will be completed in 2020, and mapping at DEVA will occur in 2021. The California Native Plant Society (CNPS) and NatureServe, the classification team, have completed the vegetation classification for these four park units, with field keys and descriptions of the vegetation types developed at the alliance level per the U.S. National Vegetation Classification (USNVC). We have compiled approximately 9,000 existing and new vegetation data records into digital databases in Microsoft Access. The resulting classification and descriptions include approximately 105 alliances and landform types, and over 240 associations. CNPS also has assisted the mapping teams during map reconnaissance visits, follow-up on interpreting vegetation patterns, and general support for the geospatial vegetation maps being produced. A variety of alliances and associations occur in the four park units. Per park, the classification represents approximately 50 alliances at LAKE, 65 at MOJA and CAMO, and 85 at DEVA. Several riparian alliances or associations that are somewhat rare (ranked globally as G3) include shrublands of Pluchea sericea, meadow associations with Distichlis spicata and Juncus cooperi, and woodland associations of Salix laevigata and Prosopis pubescens along playas, streams, and springs. Other rare to somewhat rare types (G2 to G3) include shrubland stands with Eriogonum heermannii, Buddleja utahensis, Mortonia utahensis, and Salvia funerea on rocky calcareous slopes that occur sporadically in LAKE to MOJA and DEVA. Types that are globally rare (G1) include the associations of Swallenia alexandrae on sand dunes and Hecastocleis shockleyi on rocky calcareous slopes in DEVA. Two USNVC vegetation groups hold the highest number of alliances: 1) Warm Semi-Desert Shrub & Herb Dry Wash & Colluvial Slope Group (G541) has nine alliances, and 2) Mojave Mid-Elevation Mixed Desert Scrub Group (G296) has thirteen alliances. These two groups contribute significantly to the diversity of vegetation along alluvial washes and mid-elevation transition zones.