Land-cover change in Cuba and implications for the area of distribution of a specialist's host-plant.

Changes in land cover directly affect biodiversity. Here, we assessed land-cover change in Cuba in the past 35 years and analyzed how this change may affect the distribution of Omphalea plants and Urania boisduvalii moths. We analyzed the vegetation cover of the Cuban archipelago for 1985 and 2020. We used Google Earth Engine to classify two satellite image compositions into seven cover types: forest and shrubs, mangrove, soil without vegetation cover, wetlands, pine forest, agriculture, and water bodies. We considered four different areas for quantifications of land-cover change: (1) Cuban archipelago, (2) protected areas, (3) areas of potential distribution of Omphalea, and (4) areas of potential distribution of the plant within the protected areas. We found that "forest and shrubs", which is cover type in which Omphalea populations have been reported, has increased significantly in Cuba in the past 35 years, and that most of the gained forest and shrub areas were agricultural land in the past. This same pattern was observed in the areas of potential distribution of Omphalea; whereas almost all cover types were mostly stable inside the protected areas. The transformation of agricultural areas into forest and shrubs could represent an interesting opportunity for biodiversity conservation in Cuba. Other detailed studies about biodiversity composition in areas of forest and shrubs gain would greatly benefit our understanding of the value of such areas for conservation.

Spatial-temporal dynamics of decaying stages of pelagic Sargassum spp. along shorelines in Puerto Rico using Google Earth Engine.

Coastal social-ecological systems in the Caribbean are affected by pelagic Sargassum spp. influxes and decomposition, but most satellite monitoring efforts focus on offshore waters. We developed a method to detect and spatial-temporally assess sargassum accumulations and their decaying stages along the shoreline and nearshore waters. A multi-predictor Random Forest model combining Sentinel-2 MultiSpectral Instrument reflectance bands and several vegetation, seaweed, water, and water quality indices was developed within the online Google Earth Engine platform. The model achieved 97 % overall accuracy and identified both fresh and decomposing sargassum, as well as the Sargassum-brown-tide generated from decomposing sargassum. We identified three hotspots of sargassum accumulation in La Parguera, Puerto Rico and found that sargassum was present every month in at least one of its forms during the entire time series (September 2015-January 2022). This research provides information to understand sargassum impacts and areas where mitigation efforts need to focus.

Deforestation and fires in the Brazilian Amazon from 2001 to 2020: Impacts on rainfall variability and land surface temperature.

Deforestation and fires in the Amazon are serious problems affecting climate, and land use and land cover (LULC) changes. In recent decades, the Amazon biome area has suffered constant fires and deforestation, causing severe environmental problems that considerably impact the land surface temperature (LST) and hydrological cycle. The Amazon biome lost a large forest area during this period. Thus, this study aims to analyze the deforestation and burned areas in the Amazon from 2001 to 2020, considering their impacts on rainfall variability and LST. This study used methods and procedures based on Google Earth Engine for analysis: (a) LULC evolution mapping, (b) vegetation cover change analysis using vegetation indices, (c) mapping of fires, (d) rainfall and LST analyses, and (e) analysis of climate influence and land cover on hydrological processes using the geographically weighted regression method. The results showed significant LULC changes and the main locations where fires occurred from 2001 to 2020. The years 2007 and 2010 had the most significant areas of fires in the Brazilian Amazon (233,401 km2 and 247,562 km2, respectively). The Pará and Mato Grosso states had the region's largest deforested areas (172,314 km2 and 144,128 km2, respectively). Deforestation accumulated in the 2016-2020 period is the greatest in the period analyzed (254,465 km2), 92% higher than in the 2005-2010 period and 82% higher than in the 2001-2005 period. The study also showed that deforested areas have been increasing in recent decades, and the precipitation decreased, while an increase is observed in the LST. It was also concluded that indigenous protection areas have suffered from anthropic actions.

Vegetational succession assessment in a fragment of the Brazilian Atlantic Forest.

Vegetational succession assessment is an important step for better management practices, providing relevant quantitative and qualitative information. With the advancements of remote sensing algorithms and access to data, land use and land cover (LULC) monitoring has become increasingly feasible and important for the evaluation of changes in the landscape at different spatial and temporal scales. This study aims to analyze the vegetation succession achieved by a project funded by the Brazilian Environmental Ministry (Ministério do Meio Ambiente, in Portuguese) intended to recover degraded areas. A 2014 and a 2019 LULC map was generated using high-resolution (10 cm) images. Given the great challenge of classifying high-resolution images, three classification algorithms were compared. The techniques to regenerate degraded areas were efficient to increase arboreal vegetation area by more than 30% between 2014 and 2019. Land cover and land use change monitoring is of paramount importance to strengthen sustainable practices, especially in the highly threatened Atlantic Forest biome. This study also shows that funding opportunities are essential for projects that make such actions possible, including the present research and the analysis of environmental regeneration.

Habitat loss estimation for assessing terrestrial mammalian species extinction risk: an open data framework.

Terrestrial mammals face a severe crisis of habitat loss worldwide. Therefore, assessing information on habitat loss throughout different time periods is crucial for assessing species' conservation statuses based on the IUCN Red List system. To support the national extinction risk assessment in Brazil (2016-2022), we developed a script that uses the MapBiomas Project 6.0 data source of land cover and land use (annual maps at 30 m scale) within the Google Earth Engine (GEE) platform to calculate habitat loss. We defined suitable habitats from the MapBiomas Project land cover classification for 190 mammalian taxa, according to each species range map and ecological characteristics. We considered a period of three generation lengths to assess habitat loss in accordance with the Red List assessment criteria. We used the script to estimate changes in available habitat throughout the analyzed period within the species' known ranges. The results indicated that habitat loss occurred within 94.3% of the analyzed taxa range, with the Carnivora order suffering the greatest habitat loss, followed by the Cingulata order. These analyses may be decisive for applying criteria, defining categories during the assessment of at least 17 species (9%), enriching discussions, and raising new questions for several other species. We considered the outcome of estimating habitat loss for various taxa when applying criterion A, which refers to population reduction, thus supporting more accurate inferences about past population declines.

Forecasting Weekly Dengue Cases by Integrating Google Earth Engine-Based Risk Predictor Generation and Google Colab-Based Deep Learning Modeling in Fortaleza and the Federal District, Brazil.

Efficient and accurate dengue risk prediction is an important basis for dengue prevention and control, which faces challenges, such as downloading and processing multi-source data to generate risk predictors and consuming significant time and computational resources to train and validate models locally. In this context, this study proposed a framework for dengue risk prediction by integrating big geospatial data cloud computing based on Google Earth Engine (GEE) platform and artificial intelligence modeling on the Google Colab platform. It enables defining the epidemiological calendar, delineating the predominant area of dengue transmission in cities, generating the data of risk predictors, and defining multi-date ahead prediction scenarios. We implemented the experiments based on weekly dengue cases during 2013-2020 in the Federal District and Fortaleza, Brazil to evaluate the performance of the proposed framework. Four predictors were considered, including total rainfall (Rsum), mean temperature (Tmean), mean relative humidity (RHmean), and mean normalized difference vegetation index (NDVImean). Three models (i.e., random forest (RF), long-short term memory (LSTM), and LSTM with attention mechanism (LSTM-ATT)), and two modeling scenarios (i.e., modeling with or without dengue cases) were set to implement 1- to 4-week ahead predictions. A total of 24 models were built, and the results showed in general that LSTM and LSTM-ATT models outperformed RF models; modeling could benefit from using historical dengue cases as one of the predictors, and it makes the predicted curve fluctuation more stable compared with that only using climate and environmental factors; attention mechanism could further improve the performance of LSTM models. This study provides implications for future dengue risk prediction in terms of the effectiveness of GEE-based big geospatial data processing for risk predictor generation and Google Colab-based risk modeling and presents the benefits of using historical dengue data as one of the input features and the attention mechanism for LSTM modeling.

Investigation of pre and post environmental impact of the lockdown (COVID-19) on the water quality of the Capibaribe and Tejipió rivers, Recife metropolitan region, Brazil.

The coronavirus pandemic has seriously affected human health, although some improvements on environmental indexes have temporarily occurred, due to changes on socio-cultural and economic standards. The objective of this study was to evaluate the impacts of the coronavirus and the influence of the lockdown associated with rainfall on the water quality of the Capibaribe and Tejipió rivers, Recife, Northeast Brazil, using cloud remote sensing on the Google Earth Engine (GEE) platform. The study was carried out based on eight representative images from Sentinel-2. Among the selected images, two refer to the year 2019 (before the pandemic), three refer to 2020 (during a pandemic), two from the lockdown period (2020), and one for the year 2021. The land use and land cover (LULC) and slope of the study region were determined and classified. Water turbidity data were subjected to descriptive and multivariate statistics. When analyzing the data on LULC for the riparian margin of the Capibaribe and Tejipió rivers, a low permanent preservation area was found, with a predominance of almost 100% of the urban area to which the deposition of soil particles in rivers are minimal. The results indicated that turbidity values in the water bodies varied from 6 mg. L-1 up to 40 mg. L-1. Overall, the reduction in human-based activities generated by the lockdown enabled improvements in water quality of these urban rivers.

Mapping of Land Cover with Optical Images, Supervised Algorithms, and Google Earth Engine.

Crops and ecosystems constantly change, and risks are derived from heavy rains, hurricanes, droughts, human activities, climate change, etc. This has caused additional damages with economic and social impacts. Natural phenomena have caused the loss of crop areas, which endangers food security, destruction of the habitat of species of flora and fauna, and flooding of populations, among others. To help in the solution, it is necessary to develop strategies that maximize agricultural production as well as reduce land wear, environmental impact, and contamination of water resources. The generation of crop and land-use maps is advantageous for identifying suitable crop areas and collecting precise information about the produce. In this work, a strategy is proposed to identify and map sorghum and corn crops as well as land use and land cover. Our approach uses Sentinel-2 satellite images, spectral indices for the phenological detection of vegetation and water bodies, and automatic learning methods: support vector machine, random forest, and classification and regression trees. The study area is a tropical agricultural area with water bodies located in southeastern Mexico. The study was carried out from 2017 to 2019, and considering the climate and growing seasons of the site, two seasons were created for each year. Land use was identified as: water bodies, land in recovery, urban areas, sandy areas, and tropical rainforest. The results in overall accuracy were: 0.99% for the support vector machine, 0.95% for the random forest, and 0.92% for classification and regression trees. The kappa index was: 0.99% for the support vector machine, 0.97% for the random forest, and 0.94% for classification and regression trees. The support vector machine obtained the lowest percentage of false positives and margin of error. It also acquired better results in the classification of soil types and identification of crops.

Socioeconomic inequalities and the surface heat island distribution in Santiago, Chile.

Surface urban heat islands (SUHIs) are an important socio-environmental problem associated with large cities, such as the Santiago Metropolitan Area (SMA), in Chile. Here, we analyze daytime and nighttime variations of SUHIs for each season of the year during the period 2000-2020. To evaluate socioeconomic inequities in the distribution of SUHIs, we establish statistical relationships with socioeconomic status, land price, and urban vegetation. We use the MODIS satellite images to obtain the land surface temperatures and the normalized difference vegetation index (NDVI) through the Google Earth Engine platform. The results indicate more intense SUHIs during the nighttime in the eastern sector, coinciding with higher socioeconomic status and larger green areas. This area during the day is cooler than the rest of the city. The areas with lower and middle socioeconomic status suffer more intense SUHIs (daytime and nighttime) and match poor environmental and urban qualities. These results show the high segregation of SMA. Urban planning is subordinated to land prices with a structure maintained over the study period. The lack of social-climate justice is unsustainable, and such inequalities may be exacerbated in the context of climate change. Thus, these results can contribute to the planning of the SMA.

Assessment of the vineyard water footprint by using ancillary data and EEFlux satellite images. Examples in the Chilean central zone.

The increase of vineyard's water consumption due to the Global Warming Phenomenon (GWP) has forced the winegrowers to strengthen their irrigation and water stewardship efforts, intended for maintaining this resource's long-term sustainable use. Due to water being a limited resource, implementing the Water Footprint (WF) concept in winegrapes production provides helpful information for sustainable water stewardship. Currently, an automated version of the satellite-based METRIC (Mapping Evapotranspiration with Internalized Calibration) model, the Google Earth Engine Evapotranspiration Flux (EEFlux) platform, has been suggested as an alternative to analyzing the spatial variability of an entire field's water consumption throughout the growing season. This work aimed to evaluate the potential application of the EEFlux satellite's actual evapotranspiration (ETa) products and ancillary field data to obtain the WF blue (WFb) and green (WFg) of six commercial vineyards placed in the Chilean central zone. Firstly, the reliability of the daily actual evapotranspiration data from EEFlux (ETa EEFlux) was assessed against measured ETa data, using an available database from previous studies. The results of ETa EEFlux estimations against measured ETa were impressive, presenting a root square error (RMSE) of 0.8 mm day-1. The satellite-derived crop coefficients (Kc Sat) allowed to estimate the total WF of each vineyard, in a range of 200 to 900 m3 t-1, showing an average relative error (RE) of 101%, between the satellite-based WFb (WFb Sat) and those calculated from irrigation records (WFb). These results reflected the particular conditions of each vineyard and can be considered reasonable since they were estimated from ancillary data and EEFlux products. This study provides new insights that may represent opportunities to sustainably managing the irrigation of vineyards.

Global drought monitoring with big geospatial datasets using Google Earth Engine.

Drought or dryness occurs due to the accumulative effect of certain climatological and hydrological variables over a certain period. Droughts are studied through numerically computed simple or compound indices. Vegetation condition index (VCI) is used for observing the change in vegetation that causes agricultural drought. Since the land surface temperature has minimum influence from cloud contamination and humidity in the air, so the temperature condition index (TCI) is used for studying the temperature change. Dryness or wetness of soil is a major indicator for agriculture and hydrological drought and for that purpose, the index, soil moisture condition index (SMCI), is computed. The deviation of precipitation from normal is a major cause for meteorological droughts and for that purpose, precipitation condition index (PCI) is computed. The years when the indices escalated the dryness situation to severe and extreme are pointed out in this research. Furthermore, an interactive dashboard is generated in the Google Earth Engine (GEE) for users to compute the said indices using country boundary, time period, and ecological mask of their choice: Agriculture Drought Monitoring. Apart from global results, three case studies of droughts (2002 in Australia, 2013 in Brazil, and 2019 in Thailand) computed via the dashboard are discussed in detail in this research.

Haiti has more forest than previously reported: land change 2000-2015.

Estimates of forest cover have important political, conservation, and funding implications, but methods vary greatly. Haiti has often been cited as one of the most deforested countries in the world, yet estimates of forest cover range from <1% to 33%. Here, we analyze land change for seven land cover classes (forest, shrub land, agriculture/pasture, plantation, urban/infrastructure, barren land, and water) between 2000 and 2015 using Landsat imagery (30 m resolution) in the Google Earth Engine platform. Forest cover was estimated at 26% in 2000 and 21% in 2015. Although forest cover is declining in Haiti, our quantitative analysis resulted in considerably higher forest cover than what is usually reported by local and international institutions. Our results determined that areas of forest decline were mainly converted to shrubs and mixed agriculture/pasture. An important driver of forest loss and degradation could be the high demand for charcoal, which is the principal source of cooking fuel. Our results differ from other forest cover estimates and forest reports from national and international institutions, most likely due to differences in forest definition, data sources, spatial resolution, and methods. In the case of Haiti, this work demonstrates the need for clear and functional definitions and classification methods to accurately represent land use/cover change. Regardless of how forests are defined, forest cover in Haiti will continue to decline unless corrective actions are taken to protect remaining forest patches. This can serve as a warning of the destructive land use patterns and can help us target efforts for better planning, management, and conservation.

Determining a Threshold to Delimit the Amazonian Forests from the Tree Canopy Cover 2000 GFC Data.

Open global forest cover data can be a critical component for Reducing Emissions from Deforestation and Forest Degradation (REDD+) policies. In this work, we determine the best threshold, compatible with the official Brazilian dataset, for establishing a forest mask cover within the Amazon basin for the year 2000 using the Tree Canopy Cover 2000 GFC product. We compared forest cover maps produced using several thresholds (10%, 30%, 50%, 80%, 85%, 90%, and 95%) with a forest cover map for the same year from the Brazilian Amazon Deforestation Monitoring Project (PRODES) data, produced by the National Institute for Space Research (INPE). We also compared the forest cover classifications indicated by each of these maps to 2550 independently assessed Landsat pixels for the year 2000, providing an accuracy assessment for each of these map products. We found that thresholds of 80% and 85% best matched with the PRODES data. Consequently, we recommend using an 80% threshold for the Tree Canopy Cover 2000 data for assessing forest cover in the Amazon basin.

Predicting spatiotemporal patterns of road mortality for medium-large mammals.

We modelled the spatiotemporal patterns of road mortality for seven medium-large mammals, using a roadkill dataset from Mato Grosso do Sul, Brazil (800 km of roads surveyed every two weeks, for two years). We related roadkill presence-absence along the road sections (1000 m) and across the survey dates with a collection of environmental variables, including land cover, forest cover, distance to rivers, temperature, precipitation and vegetation productivity. We further included two variables aiming to reflect the intrinsic spatial and temporal roadkill risk. Environmental variables were obtained through remote sensing and weather stations, allowing the estimate of the roadkill risk for the entire surveyed roads and survey periods. Overall, the models could explain a small fraction of the spatiotemporal patterns of roadkills (<0.23), probably due to species being habitat generalists, but still had reasonable discrimination power (AUC averaging 0.70 ±â€¯0.07). The intrinsic spatial and temporal roadkill risk were the most important variables, followed by land cover, climate and NDVI. We show that identifying spatiotemporal roadkill patterns may provide valuable information to define specific management actions focused on road sections and time periods, in complement to permanent road mitigation measures. Our approach thus offers a new insight into the understanding of road effects and how to plan and strategize monitoring and mitigation.