Mapping Forest Burn Extent from Hyperspatial Imagery Using Machine Learning

Support vector machines are shown to be highly effective in mapping burn extent from hyperspatial imagery in grasslands. Unfortunately, this pixel-based method is hampered in forested environments that have experienced low-intensity fires because unburned tree crowns obstruct the view of the surface vegetation. This obstruction causes surface fires to be misclassified as unburned. To account for misclassifying areas under tree crowns, trees surrounded by surface burn can be assumed to have been burned underneath. This effort used a mask region-based convolutional neural network (MR-CNN) and support vector machine (SVM) to determine trees and burned pixels in a post-fire forest. The output classifications of the MR-CNN and SVM were used to identify tree crowns in the image surrounded by burned surface vegetation pixels. These classifications were also used to label the pixels under the tree as being within the fire’s extent. This approach results in higher burn extent mapping accuracy by eliminating burn extent false negatives from surface burns obscured by unburned tree crowns, achieving a nine percentage point increase in burn extent mapping accuracy.

Regional unevenness of the summer warming in the continental Arctic as an indicator of natural boundaries of northern landscapes

The authors have revealed the features of summer warming in different sectors of the Russian Arctic zone in the modern period and the near future. In connection with the considered features of the summer warming within 1991—2018, the researchers present a unique analysis of the inter-decade distribution of trends in the characteristics of the natural zones surface (vegetation index, total evapotranspiration, surface temperature, albedo). Changes in climatic conditions provide prerequisites for a change in the spectral characteristics of landscape zones, especially in the central sector of the Russian Arctic zone.

Non-Destructive Fuel Volume Measurements Can Estimate Fine-Scale Biomass across Surface Fuel Types in a Frequently Burned Ecosystem

Measuring wildland fuels is at the core of fire science, but many established field methods are not useful for ecosystems characterized by complex surface vegetation. A recently developed sub-meter 3D method applied to southeastern U.S. longleaf pine (Pinus palustris) communities captures critical heterogeneity, but similar to any destructive sampling measurement, it relies on separate plots for calculating loading and consumption. In this study, we investigated how bulk density differed by 10-cm height increments among three dominant fuel types, tested predictions of consumption based on fuel type, height, and volume, and compared this with other field measurements. The bulk density changed with height for the herbaceous and woody litter fuels (p < 0.001), but live woody litter was consistent across heights (p > 0.05). Our models predicted mass well based on volume and height for herbaceous (RSE = 0.00911) and woody litter (RSE = 0.0123), while only volume was used for live woody (R2 = 0.44). These were used to estimate consumption based on our volume-mass predictions, linked pre- and post-fire plots by fuel type, and showed similar results for herbaceous and woody litter when compared to paired plots. This study illustrates an important non-destructive alternative to calculating mass and estimating fuel consumption across vertical volume distributions at fine scales.

Article Title Variations in Growing Season NDVI and Its Sensitivity to Climate Change Responses to Green Development in Mountainous Areas

The Normalized Difference Vegetation Index (NDVI) is sensitive to changes in surface vegetation cover. Research into how climate change impacts surface vegetation cover is essential to manage ecological systems and promote green development. The Western Henan Mountains, located in the transitional zone between the northern subtropical and warm temperate zones of China, is an ideal location to study the impacts of climate change on surface vegetation cover. Combining a digital elevation model (DEM) with temperature and precipitation data; and MODIS-NDVI imagery (2000∼2017) for the Western Henan Mountains, this study explores variations in the growing season NDVI and its response to climate change. Results show that there are significant changes with fluctuations in NDVI values from 2000 to 2017. NDVI increased at a growth rate of 0.027 per decade (p &lt; 0.05) overall, indicating vegetation conditions have gradually improved. Although the NDVI value showed an overall increasing trend, 19.12% of the areas showed a decreasing trend, interspersing and intersecting spatially, showing significant spatial differences. NDVI increased initially, but then decreased as a function of elevation, which was shown to be proportional to slope and independent of aspect. Variables including elevation and slope gradient are shown to provide high explanation of NDVI variability, whilst temperature is shown to have a more significant impact on NDVI than precipitation. However, vegetation responses to temperature and precipitation covaried with both slope and aspect. Positive NDVI trends were strongest at low elevations (i.e., &lt;1,100 masl), which we attribute to vegetation restoration activities. Lower NDVI values characterized gentle slopes (&lt;5°), whilst higher values were, in contrast, associated with steeper slopes (5∼10°). This study highlights the complex mechanisms and their relations governing vegetation response to climate change and should form an instructive basis for both future modeling studies investigating the response of vegetation to future global warming.

The Indirect Impact of Surface Vegetation Improvement on the Climate Response of Sand-Dust Events in Northern China

Extensive ecosystem restoration is increasingly seen as an essential practice to mitigate climate change and protect the ecological environment. However, the indirect impact of surface vegetation improvement on the regional climate, such as the climate effect of sand-dust events reduction, has never been evaluated. Here, we estimated the feedback of temperature and precipitation on the change of sand-dust events, arising from the vegetation growth with ecological restoration, using a simple theoretical framework with a series of scenario simulations based on a regional climate model (RegCM). The results showed that revegetation reduced dust emissions, with a contribution rate of approximately 40.15%. With the combined influence of ecological restoration and climate change, the cooling effect of sand-dust events strengthened with the increase in the intensity of sand-dust events, which is mainly caused by the strong absorption of shortwave radiation by the atmosphere. The response of precipitation was uncertain because of tropospheric circulation feedback and shortwave radiation absorption. Our results also indicate that changes in sand-dust events caused by vegetation restoration play important roles in shaping the future climate near the arid and semi-arid regions of northern China. The climatic effects of sand-dust events should be included in assessing ecological restoration impacts to promote sustainable development and enhance our understanding of climate change.

Thaw subsidence and frost heave caused by 2018-20 forest fires around Batagay: validation with multiple InSAR data and field observation

&lt;p&gt;The Arctic has experienced numerous fires in last year, and from June to August 2020, satellite data showed record carbon dioxide emissions from forest fires. Peatland in the Arctic contains large amounts of organic carbon, and their release into the atmosphere can create positive feedbacks for further increase of air temperature. In addition, forest fires burn the surface vegetation layer that has been acting as a heat insulator, which will accelerate the thawing of permafrost on scales of years to decades. Although the thaw depth can recover together with the recovery of surface vegetation, the massive segregated ice is not recoverable once it melted. Our study area is around the Batagay, Sakha Republic, Eastern Siberia. In June 2020, Verkhoyansk, located about 55 km west of Batagay, recorded the highest daily maximum temperature of 38.0 degrees Celcius. The Sentinel-2 optical satellite images showed a number of forest fires in 2019-20. We detected the surface deformation signals at each fire site with the remote-sensing method called InSAR (Interferometric Synthetic Aperture Radar). Also, we conducted a field observation in September 2019 for validations: 1) installed a soil thermometer and soil moisture meter; 2) established a reference point for leveling and first survey; 3) measured the thawing depth with a frost probe.&lt;/p&gt;&lt;p&gt;&amp;#160;For seasonal ground deformations immediately after the fire, we mainly analyzed Sentinel-1 images. Sentinel-1 is the ESA's C-band SAR satellite, which has a short imaging interval of 12 days. As the short wavelength, vegetation changes lost coherence, and some pairs failed to detect ground deformation signals immediately after the fire. However, after the end of September, we detected displacements toward the satellite line-of-sight direction at all the fire sites. It indicates uplift signals due presumably to frost heave at the fire scar. For long-term deformations over one year, we used ALOS2 imaged derived by JAXA's L band SAR satellite. In the previous studies in Alaska, the ground deformation signal immediately after a fire could not be detected due to the coherence loss in the pairs derived from pre-fire and post-fire SAR images. Indeed, we could not detect deformation signals at the fire scars from the June pairs derived before and after the fire. However, the January pairs and March pairs, both of which were acquired before and after the fire, showed relatively high coherence even in the fire scar and indicated clear subsidence signals by as much as 15 cm. We interpret that, because the studied Verkhoyansk Basin is very dry and has little snow cover, the microwaves could penetrate the snow layer, which allowed us to detect deformation signals even in winter. Yanagiya and Furuya (2020) validated the consistency of the winter uplift signal for the 2014 fire site. We also analyzed the SM1 high spatial resolution mode (3 m) ALOS2 InSAR to investigate the specific ground deformation at each fire site.&lt;/p&gt;

Reduction in Human Activity can Enhance the Urban Heat Island: Insights from the COVID-19 Lockdown

The COVID-19 lockdowns drastically reduced human activity, emulating a controlled experiment on human-land-atmosphere coupling. Here, using a fusion of satellite and reanalysis products, we examine this coupling during the lockdown in the Indo-Gangetic Basin, one of the world’s most populated and polluted regions. During the lockdown, the reduction (&gt;10%) in columnar air pollution, expected to increase incoming solar radiation, was counteracted by a ~30% enhancement in cloud cover, causing little change in available energy at the surface. More importantly, the delay in winter crop harvesting increased surface vegetation cover, causing almost half the regional cooling (of -3.9 K) via evapotranspiration. Since this cooling was higher for rural areas, the daytime surface urban heat island (SUHI) intensity increased (by 0.20 to 0.41 K) during a period of reduced human activity. Our study provides strong observational evidence of the influence of agricultural activity on both urban and rural climate in this region.

Single Column Model Simulations of Icing Conditions in Northern Sweden: Sensitivity to Surface Model Land Use Representation

In-cloud ice mass accretion on wind turbines is a common challenge that is faced by energy companies operating in cold climates. On-shore wind farms in Scandinavia are often located in regions near patches of forest, the heterogeneity length scales of which are often less than the resolution of many numerical weather prediction (NWP) models. The representation of these forests—including the cloud water response to surface roughness and albedo effects that are related to them—must therefore be parameterized in NWP models used as meteorological input in ice prediction systems, resulting in an uncertainty that is poorly understood and, to the present date, not quantified. The sensitivity of ice accretion forecasts to the subgrid representation of forests is examined in this study. A single column version of the HARMONIE-AROME three-dimensional (3D) NWP model is used to determine the sensitivity of the forecast of ice accretion on wind turbines to the subgrid forest fraction. Single column simulations of a variety of icing cases at a location in northern Sweden were examined in order to investigate the impact of vegetation cover on ice accretion in varying levels of solar insolation and wind magnitudes. In mid-winter cases, the wind speed response to surface roughness was the primary driver of the vegetation effect on ice accretion. In autumn cases, the cloud water response to surface albedo effects plays a secondary role in the impact of in-cloud ice accretion, with the wind response to surface roughness remaining the primary driver for the surface vegetation impact on icing. Two different surface boundary layer (SBL) forest canopy subgrid parameterizations were tested in this study that feature different methods for calculating near-surface profiles of wind, temperature, and moisture, with the ice mass accretion again following the wind response to surface vegetation between both of these schemes.