A global land cover training dataset from 1984 to 2020.

State-of-the-art cloud computing platforms such as Google Earth Engine (GEE) enable regional-to-global land cover and land cover change mapping with machine learning algorithms. However, collection of high-quality training data, which is necessary for accurate land cover mapping, remains costly and labor-intensive. To address this need, we created a global database of nearly 2 million training units spanning the period from 1984 to 2020 for seven primary and nine secondary land cover classes. Our training data collection approach leveraged GEE and machine learning algorithms to ensure data quality and biogeographic representation. We sampled the spectral-temporal feature space from Landsat imagery to efficiently allocate training data across global ecoregions and incorporated publicly available and collaborator-provided datasets to our database. To reflect the underlying regional class distribution and post-disturbance landscapes, we strategically augmented the database. We used a machine learning-based cross-validation procedure to remove potentially mis-labeled training units. Our training database is relevant for a wide array of studies such as land cover change, agriculture, forestry, hydrology, urban development, among many others.

A comprehensive assessment of mangrove species and carbon stock on Pohnpei, Micronesia.

Mangrove forests are the most important ecosystems on Pohnpei Island, Federated States of Micronesia, as the island communities of the central Pacific rely on the forests for many essential services including protection from sea-level rise that is occurring at a greater pace than the global average. As part of a multi-component assessment to evaluate vulnerabilities of mangrove forests on Pohnpei, mangrove forests were mapped at two points in time: 1983 and 2018. In 2018, the island had 6,426 ha of mangrove forest. Change analysis indicated a slight (0.76%) increase of mangrove area between 1983 and 2018, contrasting with global mangrove area declines. Forest structure and aboveground carbon (AGC) stocks were inventoried using a systematic sampling of field survey plots and extrapolated to the island using k-nearest neighbor and random forest species models. A gridded or wall to wall approach is suggested when possible for defining carbon stocks of a large area due to high variability seen in our data. The k-nearest neighbor model performed better than random forest models to map species dominance in these forests. Mean AGC was 167 ± 11 MgC ha-1, which is greater than the global average of mangroves (115 ± 7 MgC ha-1) but within their global range (37-255 MgC ha-1) Kauffman et al. (2020). In 2018, Pohnpei mangroves contained over 1.07 million MgC in AGC pools. By assigning the mean AGC stock per species per area to the map, carbon stock distributions were visualized spatially, allowing future conservation efforts to be directed to carbon dense stands.

Carbon loss and removal due to forest disturbance and regeneration in the Amazon.

The forest carbon flux is the difference between the total carbon loss from deforestation, forest degradation, and natural disturbance and removal of atmospheric CO2 due to photosynthetic activity. The Amazon rainforest accounts for approximately a quarter of global emissions from land use change, due in part to its' immense size, carbon storage, and recent history of land use change. Large area estimates of carbon exchange in forests are highly uncertain, however, which reflects the pervasive challenges in estimating carbon flux parameters, such as disturbance area and forest carbon pools. In this study, we use a new dataset with characterized uncertainty on deforestation, degradation, and natural disturbances in the Amazon Ecoregion to estimate carbon loss from disturbance and removals from regeneration at biennial intervals from 1996 to 2017. Using the gain-loss approach to estimating carbon flux in a Monte Carlo analysis we found that carbon loss from degradation and deforestation averaged 0.23 (±0.09) Pg C biennium-1 and 0.34 (±0.16) Pg C biennium-1, respectively. While deforestation contributed the most to carbon loss overall, there were two biennial periods in which degradation and natural disturbance resulted in more carbon loss. Regeneration partially offset these emissions, but our results show that loss is occurring much more rapidly than removal, resulting in a total net carbon loss of 4.86 to 5.32 Pg C over the study period. With the compounding effect of drought and fires in addition to continued deforestation it appears certain that forest disturbance in the Amazon will continue to be a significant factor in the terrestrial carbon cycle.

Ongoing forest disturbance in Guatemala's protected areas.

Protected areas in Guatemala provide habitat for diverse tropical ecosystems, contain ancient archeological sites, sequester carbon, and support economic activity through eco-tourism. However, many of the forests in these protected areas have been converted to other uses or degraded by human activity, and therefore are considered "paper parks". In this study, we analyzed time series of satellite data to monitor deforestation, degradation, and natural disturbance throughout Guatemala from 2000 to 2017. A recently developed methodology, Continuous Degradation Detection (CODED), was used to detect forest disturbances of varying size and magnitude. Through sample-based statistical inference, we estimated that 854 137 ha (± 83 133 ha) were deforested and 1 012 947 ha (±139 512 ha) of forest was disturbed but not converted during our study period. Forest disturbance in protected areas ranged from under 1% of a park's area to over 95%. Our estimate of the extent of deforestation is similar to previous studies, however, degradation and natural disturbance affect a larger area. These results suggest that the total amount of forest disturbance can be significantly underestimated if degradation and natural disturbance are not taken into account. As a consequence, we found that the protected areas of Guatemala are more affected by disturbance than previously realized.

Satellite-based estimates reveal widespread forest degradation in the Amazon.

Anthropogenic and natural forest disturbance cause ecological damage and carbon emissions. Forest disturbance in the Amazon occurs in the form of deforestation (conversion of forest to non-forest land covers), degradation from the extraction of forest resources, and destruction from natural events. The crucial role of the Amazon rainforest in the hydrologic cycle has even led to the speculation of a disturbance "tipping point" leading to a collapse of the tropical ecosystem. Here we use time series analysis of Landsat data to map deforestation, degradation, and natural disturbance in the Amazon Ecoregion from 1995 to 2017. The map was used to stratify the study area for selection of sample units that were assigned reference labels based on their land cover and disturbance history. An unbiased statistical estimator was applied to the sample of reference observations to obtain estimates of area and uncertainty at biennial time intervals. We show that degradation and natural disturbance, largely during periods of severe drought, have affected as much of the forest area in the Amazon Ecoregion as deforestation from 1995 to 2017. Consequently, an estimated 17% (1,036,800 ± 24,800 km2 , 95% confidence interval) of the original forest area has been disturbed as of 2017. Our results suggest that the area of disturbed forest in the Amazon is 44%-60% more than previously realized, indicating an unaccounted for source of carbon emissions and pervasive damage to forest ecosystems.