Silvilaser 2019 is one of the best conferences for networking and getting updated on what scientists face when monitoring the world forests. The 2019 team of Silvilaser keynote speakers will give us their personal views on some of these challenges and achievements. A special group of invited speakers has already confirmed their participation and they will share with us their personal experiences and opinions on how active remote sensors have evolved and how these technologies are promoting the quick development of the ways we monitor and study natural resources in our planet. See what they are preparing to bring to us.
I provide an overview of how unmanned aerial vehicles fit into a framework for ecological mapping in tropical forests, discussing potential synergism among new satellites (e.g., Planet, GEDI), airborne systems, and drone platforms. I will describe the University of Florida’s GatorEye Unmanned Flying Laboratory, a newly developed system integrating simultaneous co-aligned LiDAR, upward and downward facing hyperspectral, radiometric thermal, and visual sensors. Following a discussion of hardware and technical specifications of the sensors, and software post-processing and integration approaches, I will discuss how they contribute to ecological mapping of different key forest parameters and how their combination through fusion algorithms results in significantly greater ecological insights. To provide context and examples, I will describe a variety of projects, associated field campaigns, and processing approaches that use the GatorEye, as well as other platforms. I will conclude with a discussion of future research directions, both in terms of hardware and processing algorithms, but especially questions related to conservation biology, ecology, and archeology.
Estimates of biomass and carbon for the Amazonian biome has been the object of several studies. The studies have used different approaches, methodologies, and data. INPE's Earth Science Center has supported research meant to improve the accuracy of these estimates. With resources from the Amazon Fund, 950 non-overlapping LiDAR airborne strips (300 m wide by 12.5 km long, 375 ha each), and 50 hyperspectral transects were obtained and processed to generate a new set of estimates for biomass and carbon in the Amazon. Some of these transects were targeted to cover areas in which forest inventory sample plots had already been measured. This talk will present the percentage of forest types covered by the assessment and comment about the allometric equations used, the biomass and carbon estimation at the field plot level, the selection of LiDAR metrics and model fitting, and the use of complementary satellite images to extrapolate estimates to the whole Amazonian region.
As a sequel to the story told at the 2012 Vancouver Silvilaser conference, the presentation now shifts towards the nomadic on-site development of the LAStools suite for point cloud processing. Subtly rich in human and technical aspects, the saga spans seven years of experiences with users in different parts of the world. The challenge of merging rebel flexibility with lovely perfectionism spices the story. The talk covers some key events that have spurred the development of new tools and features, and how these outcomes have supported and inspired others. The point of view is that of a scientific-minded, nature-loving vagabond who pursues a make-it-up-as-you-go mashup of technical challenges and personal passions in the complex landscapes of a world described by every denser and richer point clouds. The narrative will also paint a picture of where this odyssey might be heading and how forest managers, educators, environmental scientist, and remote sensing researchers may benefit from or can participate in its future.
Droughts associated with El Niño events can lead to leaf loss, tree mortality and lower productivity in tropical rain forests, but what remains unclear is how and why drought impacts vary across tropical landscapes. We used repeat airborne LiDAR surveys conducted before and after the strong El Niño event of 2015-16 in Malaysian Borneo to produce high-resolution maps of canopy height change across 25,000 ha of a human-modiﬁed tropical landscape. Using a combination of LiDAR surveys, topographic models, field measurements of aboveground biomass change and microclimate, we found that tall forest canopies on hilltops shrunk by an average of 0.6 ± 0.3 m following the El Niño whereas regenerating forests in valleys grew 1 ± 0.2 m. Long-term tree census and Leaf Area Index data revealed that canopy loss in canopy height was primarily driven by leaf shedding rather than tree mortality, which is consistent with a short-term physiological response owing to increased water pressure deficit. This study has uncovered environmental controls on forest canopy dynamics over unprecedented scales and demonstrated the power of repeat airborne LiDAR to understand forest responses to extreme climatic events.
Deforestation rates in Brazil have decreased by about 70% since 2004 but forest degradation processes including logging, fire, and fragmentation continue to change forest structure and deplete carbon stocks. Airborne lidar remote sensing provides a window on forest changes whether from natural processes such as droughts, deliberate management such as reduced impact logging, or unmanaged human induced destruction through predatory logging and fire. In this talk, I will present examples from a wide range of forest sites collected by the Sustainable Landscapes Brazil program, a partnership operated by the US Forest Service and EMBRAPA. As a sidelight, I will also tell how Sustainable Landscapes Brazil engaged a large community of Brazilians practitioners and researchers in forest studies and management. I will highlight how lidar helps us to understand the ecological processes in dynamic frontier forests and examine what we can learn from lidar to manage the future of tropical forests.
For over 50 years NASA, along with other international space agencies, has supported important scientific and policy issues related to the present and future states of the Earth’s carbon and water cycles, its climate, its habitat suitability, and other ecosystem services using a constellation of Earth-orbiting satellites. The most direct and accurate way of obtaining vegetation three-dimensional structure and its above-ground carbon content is through lidar remote sensing. The Global Ecosystem Dynamics was selected as part of NASA's Earth System Science Pathfinder (ESSP) Earth Ventures 2 (EV-2) competition. GEDI was launched from Cape Canaveral, Florida in the Dragon capsule of SpaceX CRS-16 on board of a Falcon 9 rocket and subsequently installed on the Japanese Experiment Module-Exposed Facility (JEM-EF) in December of 2018. Using its three lasers, GEDI provides billions of observations of vegetation vertical structure over the Earth’s temperate and tropical forests. GEDI lidar observations are the first set of space-borne measurements from an instrument specifically designed to measure vegetation structure. In this talk, I provide the GEDI mission scientific goals and objectives, explain its rationale within a context of carbon balance and biodiversity and next I present the GEDI lidar instrument, its measurement capabilities, and some early results.
Among the variety of remote sensing (RS) data types, LiDAR data have been shown to be strongly correlated with forest attributes such as volume and biomass. The strong correlation makes it possible to treat LiDAR-based predictions as pseudo-field reference data, which in turn can be used to train models linking forest attributes with other types of remote sensing data. Such a setup allows for forest surveys in remote areas at minimal fieldwork effort. Thus, the use of sample LiDAR data in an intermediate step of model-based large-area forest surveying offers many possibilities. But there are also pitfalls. An obvious one is that all sources of uncertainty are not accounted for when overall uncertainty metrics are computed. For example, if the uncertainty in the LiDAR to field data modeling step is ignored the overall variance of mean biomass estimators may be substantially underestimated; a recent study suggests that the underestimation may amount to 70%. With hierarchical model-based (HMB) inference it is possible to handle all sources of modeling uncertainty in an appropriate way in this type of surveys.
An overview will be given on the evolution of Terrestrial Laser Scanning, point cloud processing techniques and tools for getting accurate tree-wise information, focusing on how terrestrial laser surveys are now scalable for large operational applications, thanks to the current stage of Mobile Laser Scanning (MLS) systems. MLS has shown its potential for replacing the traditional forest inventory strategies by applying (semi) automated approaches based on cutting edge technology, from fieldwork to data analysis. By mixing portable LiDAR sensors and state of the art cloud co-registration algorithms, fast and accurate MLS surveys are now possible regardless of auxiliary GNSS. The current possibilities of MLS applied to both planted and natural forests monitoring will be discussed in depth, bringing up the main challenges and opportunities that arise from it on the perspective of precision forestry, LiDAR-based technologies, data processing techniques, and software development.