Large Spatial Extents Research Articles

Topographic Drivers of Soil Moisture Across a Large Sensor Network in the Southern Appalachian Mountains (USA)

AbstractUnderstanding the distribution of soil moisture is notoriously difficult in topographically complex regions that are subject to both large‐scale climate gradients and fine‐scale effects of terrain, vegetation, and soil structure. Remote sensing approaches capture large‐scale moisture patterns but are limited in spatial and temporal resolution, while commercial field sensors remain too expensive to deploy intensively over large spatial extents. Here, we demonstrate the use of low‐cost (<20 USD) custom sensors to create a large monitoring network of surficial (0–15 cm depth) volumetric soil moisture content (VMC) across Great Smoky Mountains National Park (GSMNP) (NC, TN, USA). In laboratory tests, temperature‐calibrated VMC values approached the accuracy of commercial probes. We deployed over 80 sensors across multiple watersheds, topographic positions, and a 1,800‐m elevation gradient, and created hierarchical models to understand associations of VMC with spatial (30‐m resolution) and temporal (daily) variables related to water supply and demand. Elevation had the strongest association with VMC, with a fivefold increase across the gradient reflecting 1.5‐fold changes in both (increased) precipitation and (decreased) evapotranspiration; slope angle was a strong mediating factor. Common proxies for moisture including topographic convergence index were not associated with VMC, likely due to limited contributions of surface drainage to local water balance. Our model predicted daily VMC of a set of validation sensors with a root mean square error of 4.8%, which may be improved by site‐specific field calibration. Our study indicates that spatially extensive, field‐based soil moisture networks are practical, accurate, and an important component of regional environmental monitoring.

Linking ecosystem processes to consumer growth rates: gross primary productivity as a driver of freshwater fish somatic growth in a resource-limited river

Individual growth can exert strong control on population dynamics but is constrained by resource acquisition rates. Difficulty in accurately quantifying resource availability over large spatial extents and at high temporal frequencies often limits attempts to understand the extent to which resources limit individual growth. Daily estimates of stream metabolism, including gross primary productivity (GPP), are increasingly available but have not, to our knowledge, been linked to fish growth. Here we examine how environmental variables such as GPP, water temperature, turbidity, and high-flow releases from a dam are linked to spatiotemporal variation in the growth of flannelmouth sucker ( Catostomus latipinnis) in the Colorado River within the Grand Canyon. We fit state-space growth models to 6 years of mark–recapture data collected in four river reaches spanning 300 river kilometers. Consistent with past research in this system, we find that all four environmental variables influence growth in length of a native primary consumer fish. GPP and temperature have a positive influence on growth, while turbidity and high-flow events have a negative influence. Water temperature is the dominant driver of spatiotemporal variation in growth, while the link between high-frequency GPP and fish growth is relatively novel. Fish growth is likely to be linked to stream metabolism in other systems where overall productivity, not the quality of primary producers, limits the food webs that support fish growth.

Automatic detection of migrating soaring bird flocks using weather radars by deep learning

Abstract The use of weather radars to detect and distinguish between different biological patterns greatly improves our understanding of aeroecology and its consequences for our lives. Importantly, it allows us to quantify passerine bird migration at different scales. Yet, no algorithm to detect soaring bird flocks in weather radar is available, precluding our ability to study this type of migration over large spatial scales. We developed the first automatic algorithm for detecting the migration of flocks of soaring birds, an important bio‐flow phenomenon involving many millions of birds that travel across large spatial extents, with implications for risk of bird‐aircraft collisions. The algorithm was developed with a deep learning network for semantic segmentation using U‐Net architecture. We tested several models with different weather radar products and with image sequences for flock movement identification. The best model includes the radial velocity product and a sequence of two previous images. It identifies 93% of soaring bird flocks that were tagged by a human on the radar image, with a false discovery of less than 20%. Large birds such as those detected by the algorithm pose a serious risk for flight safety of civilian and military transportation and therefore the application of this algorithm can substantially reduce bird‐strikes, leading to reduced financial losses and threats to human lives. In addition, it can help overcome one of the main challenges in the study of bird migration by automatically and continuously detecting flocks of large birds over wide spatial scales without the need to equip the birds with tracking devices, unravelling the abundance, timing, spatial flyways, seasonal trends and influences of environmental conditions on the migration of bird flocks.

Using machine learning to achieve simultaneous, georeferenced surveys of fish and benthic communities on shallow coral reefs

AbstractSurveying coastal systems to estimate distribution and abundance of fish and benthic organisms is labor‐intensive, often resulting in spatially limited data that are difficult to scale up to an entire reef or island. We developed a method that leverages the automation of a machine learning platform, CoralNet, to efficiently and cost‐effectively allow a single observer to simultaneously generate georeferenced data on abundances of fish and benthic taxa over large areas in shallow coastal environments. Briefly, a researcher conducts a fish survey while snorkeling on the surface and towing a float equipped with a handheld GPS and a downward‐facing GoPro, passively taking ~ 10 photographs per meter of benthos. Photographs and surveys are later georeferenced and photographs are automatically annotated by CoralNet. We found that this method provides similar biomass and density values for common fishes as traditional scuba‐based fish counts on fixed transects, with the advantage of covering a larger area. Our CoralNet validation determined that while photographs automatically annotated by CoralNet are less accurate than photographs annotated by humans at the level of a single image, the automated approach provides comparable or better estimations of the percent cover of the benthic substrates at the level of a minute of survey (~ 50 m2 of reef) due to the volume of photographs that can be automatically annotated, providing greater spatial coverage of the site. This method can be used in a variety of shallow systems and is particularly advantageous when spatially explicit data or surveys of large spatial extents are necessary.

Vegetation structure from LiDAR explains the local richness of birds across Denmark.

Classic ecological research into the determinants of biodiversity patterns emphasised the important role of three-dimensional (3D) vegetation heterogeneity. Yet, measuring vegetation structure across large areas has historically been difficult. A growing focus on large-scale research questions has caused local vegetation heterogeneity to be overlooked compared with more readily accessible habitat metrics from, for example, land cover maps. Using newly available 3D vegetation data, we investigated the relative importance of habitat and vegetation heterogeneity for explaining patterns of bird species richness and composition across Denmark (42,394 km2 ). We used standardised, repeated point counts of birds conducted by volunteers across Denmark alongside metrics of habitat availability from land-cover maps and vegetation structure from rasterised LiDAR data (10 m resolution). We used random forest models to relate species richness to environmental features and considered trait-specific responses by grouping species by nesting behaviour, habitat preference and primary lifestyle. Finally, we evaluated the role of habitat and vegetation heterogeneity metrics in explaining local bird assemblage composition. Overall, vegetation structure was equally as important as habitat availability for explaining bird richness patterns. However, we did not find a consistent positive relationship between species richness and habitat or vegetation heterogeneity; instead, functional groups displayed individual responses to habitat features. Meanwhile, habitat availability had the strongest correlation with the patterns of bird assemblage composition. Our results show how LiDAR and land cover data complement one another to provide insights into different facets of biodiversity patterns and demonstrate the potential of combining remote sensing and structured citizen science programmes for biodiversity research. With the growing coverage of LiDAR surveys, we are witnessing a revolution of highly detailed 3D data that will allow us to integrate vegetation heterogeneity into studies at large spatial extents and advance our understanding of species' physical niches.

RSARE: A physically-based vegetation index for estimating wheat green LAI to mitigate the impact of leaf chlorophyll content and residue-soil background

The green leaf area index (LAI) is an important structural parameter that can be used for monitoring plant transpiration and carbon cycling over large spatial extents. LAI is also critical for understanding biophysical processes and predicting crop productivity. However, the retrieval of wheat green LAI based on canopy reflectance is always affected by the background material. For example, in the Yangtze River Basin of China, rice residues are typically returned to rice–wheat rotation fields. Thus the background for observing wheat canopy varies from a mixture of traditional wheat-soil components to wheat-residue or wheat-residue-soil, posing a significant challenge for accurate LAI estimation at the early stages of crop growth. Furthermore, when current vegetation indices (VIs) are used to estimate LAI, leaf chlorophyll content (LCC) and field moisture are two additional confounding factors. To resolve these issues, we propose a novel residue-soil adjusted red edge difference index (RSARE), which draws on concepts from a simple physical algorithm, i.e., linear spectral mixture analysis (LSMA). The results of field, satellite and modeling experiments demonstrate that the RSARE-LAI model is generally not sensitive to LCC, or to variations in residue-soil and traditional soil backgrounds, and estimates LAI with high accuracy (RMSEval = 0.55, RRMSEval=20.71%). In comparison to traditional VI-LAI maps, RASRE-LAI maps are less sensitive to confounding factors associated with the background components, field moisture, LCC, and both inter-annual and geographical differences. This novel residue-soil background-resistant RSARE particularly improves the accuracy of wheat LAI estimation at low LAI levels, thus facilitating large-scale mapping of LAI in the early growing season.

Using just a canopy height model to obtain lidar-level accuracy in 3D forest canopy shortwave transmissivity estimates

This study presents a new model for calculating canopy shortwave radiation transmissivity from synthetic hemispheric images using only information contained within a canopy height model (CHM) – CanopyHeightModel2Radiation (C2R). The enhanced version calculates synthetic hemispherical images based on the geometric arrangement of the surrounding canopy while applying a statistical correction for canopy transmissivity using canopy thickness and tree species leaf area. The simple input data and statistical correction make this model suitable for estimating canopy transmissivity across large spatial extents typical of land surface models for which canopy transmissivity or radiation is a primary input variable. Performance of C2R-enhanced is assessed against hemispherical photographs, and compared to a basic version of C2R without transmissivity correction, and two versions of a Lidar2Radiation model (L2R-enhanced, L2R-basic) with either a basic representation of canopy structure or an enhanced representation including trunks and branches within tree crowns. The two enhanced models (L2R-enhanced and C2R-enhanced) perform best compared to hemispherical photographs, while the L2R-basic and C2R-basic models over- and underestimate canopy transmissivity, respectively. At 1-meter and 10-minute resolution, the two enhanced models perform similarly, but exact timing and location of transmissivity controlled by canopy structure is better represented in the physically explicit L2R-enhanced model. Across hourly and 25 × 25 m grid-averaged scales, both enhanced models achieve similar estimates of canopy transmissivity. Based on these results, it is recommended that the purely physically-based representation in the L2R-enhanced model is used when estimates of canopy transmissivity at high spatial and temporal (meter and minute) resolutions are necessary, while the computationally more efficient C2R-enhanced model is used when calculating canopy transmissivity within spatially aggregated grid cells, for example, as input into coarser-resolution land surface models. Incorporating C2R-enhanced into existing forest energy balance models creates exciting opportunities for investigating forest structure changes on forest hydrology and ecosystems across previously impossible spatial extents.

Using aerial LiDAR to assess regional availability of potential habitat for a conservation dependent forest bird

Remotely-sensed data can inform conservation efforts that target forest wildlife, however, few spatial data products are able to quantify fine-scale aspects of structural variation within forests. Increased availability of Light Detection and Ranging (LiDAR) datasets that cover broad spatial extents and ownership types (e.g., entire states) provide useful information regarding canopy and understory structure within forested landscapes. The fusion of LiDAR data with field-based species surveys can advance our understanding of species-habitat relationships and improve the effectiveness of conservation programs to recover habitat-limited species. The Golden-winged Warbler (Vermivora chrysoptera) is a forest-dependent songbird that nests in structurally-complex young forest across eastern North America. As with many early-successional obligates, this species has been declining for decades due, in part, to the steady loss of young forest/shrubland nesting habitat. Although conservation programs have begun restoring Golden-winged Warbler habitat, these efforts are currently limited by the inability to identify existing habitat across large spatial extents and diverse ownership patterns. Recent availability of state-wide LiDAR data for Pennsylvania provides an opportunity to overcome this limitation. From 2019 to 20, we surveyed for Golden-winged Warblers and structural vegetation at 837 sites across six forest blocks in eastern Pennsylvania. We combined these data with LiDAR derived forest structural metrics to develop statistical models to predict patterns of occupancy. Golden-winged Warbler occupancy probability was explained by models containing several LiDAR-derived structural metrics (e.g., percentage of first returns between 1 and 5 m in height, structural complexity, etc.). Moreover, models fit with LiDAR-derived covariates predicted occupancy much better than those using only field-measured vegetation covariates (ΔAICc = 53.27). Mapped predictions of Golden-winged Warbler occupancy revealed potential habitat (especially regenerating timber harvests) on both private and public lands. These results demonstrate the efficacy of LiDAR for modeling forest bird habitat associations, and how such data sources can provide a valuable tool for conservation planning.

Rapid changes in functional trait expression and decomposition following high severity fire and experimental warming

Uncharacteristically severe wildfires are occurring at higher frequency, across larger spatial extents, and in new seasons in many parts of the globe. At the same time, climate change is elevating temperatures and altering precipitation patterns. High severity fires have the potential to produce shifts in ecosystem type and function in communities that are adapted to low severity fire via changes in community composition, functional trait values, and nutrient cycling processes. However, interactive effects between climate warming and fire severity on community composition, trait values, and ecosystem functioning are rarely studied and poorly understood. We assessed the impact of experimental warming via open top warming chambers across a burn severity gradient on the understory plant community of a Ponderosa pine (Pinus ponderosa) forest in northern Arizona, USA. Specifically, we examined community composition, three plant functional traits (specific leaf area (SLA), leaf dry matter content (LDMC), and plant height) and one ecosystem function (decomposition rate) in the first post-fire growing season. High burn severity produced significant shifts in community composition. The combination of experimental warming and high burn severity significantly decreased community weighted mean LDMC and plant height, and increased decomposition rate. Our study demonstrates significant and rapid responses of community composition, trait expression, and ecosystem function in response to burn severity, experimental warming, and their interaction. This suggests that ecosystems experiencing atypically severe fire under future climate and wildfire conditions may recover in fundamentally different ways than in the past, favoring a different suite of species and traits, with altered ecosystem function.

Exploring the Influence of Input Feature Space on CNN‐Based Geomorphic Feature Extraction From Digital Terrain Data

AbstractMany studies of Earth surface processes and landscape evolution rely on having accurate and extensive data sets of surficial geologic units and landforms. Automated extraction of geomorphic features using deep learning provides an objective way to consistently map landforms over large spatial extents. However, there is no consensus on the optimal input feature space for such analyses. We explore the impact of input feature space for extracting geomorphic features from land surface parameters (LSPs) derived from digital terrain models (DTMs) using convolutional neural network (CNN)‐based semantic segmentation deep learning. We compare four input feature space configurations: (a) a three‐layer composite consisting of a topographic position index (TPI) calculated using a 50 m radius circular window, square root of topographic slope, and TPI calculated using an annulus with a 2 m inner radius and 10 m outer radius, (b) a single illuminating position hillshade, (c) a multidirectional hillshade, and (d) a slopeshade. We test each feature space input using three deep learning algorithms and four use cases: two with natural features and two with anthropogenic features. The three‐layer composite generally provided lower overall losses for the training samples, a higher F1‐score for the withheld validation data, and better performance for generalizing to withheld testing data from a new geographic extent. Results suggest that CNN‐based deep learning for mapping geomorphic features or landforms from LSPs is sensitive to input feature space. Given the large number of LSPs that can be derived from DTM data and the variety of geomorphic mapping tasks that can be undertaken using CNN‐based methods, we argue that additional research focused on feature space considerations is needed and suggest future research directions. We also suggest that the three‐layer composite implemented here can offer better performance in comparison to using hillshades or other common terrain visualization surfaces and is, thus, worth considering for different mapping and feature extraction tasks.

Relating weather radar data to migrating waterfowl abundance in the Rainwater Basin of Nebraska

AbstractWaterfowl migrations are large‐scale events that involve millions of birds moving over broad geographic extents, which make them difficult to quantify and study. Historically, wildlife managers have relied mostly on field surveys, such as visual counts from the ground or air that sample at small spatial or temporal extents, or both. Combining field surveys with remote sensing data comprehensively collected over large spatial extents at high temporal frequency may improve the study of migrating waterfowl distributions. We tested the strength of the relationship between broad‐scale weather surveillance radar data and fine‐scale field surveys of waterfowl abundance at wetlands within the Rainwater Basin of Nebraska, USA, from 2017–2019. Radar reflectivity of waterfowl at the peak exodus of evening flights was positively correlated with diurnal waterfowl count data, although there was unexplained variation in the relationship. The association was also very similar across various time scales ranging from daily to monthly averages of waterfowl abundance. We suggest that human‐based ground surveys can calibrate and leverage more comprehensive remote sensing data to get a broad understanding of waterfowl distributions during migration. Several confounding factors, such as spatial displacement between radar and survey observation, individual variation in orientation and body size, and identification of avian species sampled by radar, remain on the path to improving radar‐based waterfowl estimates.

Compositional breakpoints of freshwater plant communities across continents

Unravelling patterns and mechanisms of biogeographical transitions is crucial if we are to understand compositional gradients at large spatial extents, but no studies have thus far examined breakpoints in community composition of freshwater plants across continents. Using a dataset of almost 500 observations of lake plant community composition from six continents, we examined, for the first time, if such breakpoints in geographical space exist for freshwater plants and how well a suite of ecological factors (including climatic and local environmental variables) can explain transitions in community composition from the subtropics to the poles. Our combination of multivariate regression tree (MRT) analysis and k-means partitioning suggests that the most abrupt breakpoint exists between temperate to boreal regions on the one hand and freshwater plant communities harbouring mainly subtropical or Mediterranean assemblages on the other. The spatially structured variation in current climatic conditions is the most likely candidate for controlling these latitudinal patterns, although one cannot rule out joint effects of eco-evolutionary constraints in the harsher high-latitude environments and post-glacial migration lags after Pleistocene Ice Ages. Overall, our study supports the foundations of global regionalisation for freshwater plants and anticipates further biogeographical research on freshwater plant communities once datasets have been harmonised for conducting large-scale spatial analyses.

Beaver pond identification from multi-temporal and multi- sourced remote sensing data

ABSTRACT The maintenance and restoration of wetland habitat is a priority conservation action for most waterfowl and other wetland-dependent species in North America. Despite much progress in targeting habitat management in staging and wintering areas, methods to identify and target high-quality breeding habitats that result in the greatest potential for wildlife are still required. This is particularly true for species that breed in remote, inaccessible areas such as the American black duck, an intensively managed game bird in Eastern North America. Although evidence suggests that black ducks prefer productive, nutrient-rich waterbodies, such as beaver ponds, information about the distribution and quality of these habitats across the vast boreal forest is lacking with accurate identification remaining a challenge. Continuing advancements in remote sensing technologies that provide spatially extensive and temporally repeated information are particularly useful in meeting this information gap. In this study, we used multi-source remotely sensed information and a fuzzy analytical hierarchy process to map the spatial distribution of beaver ponds in Ontario. The use of multi-source data, including a Digital Elevation Model, a Sentinel-2 Multi-Spectral Image, and RadarSat 2 Polarimetric data, enabled us to identify individual beaver ponds on the landscape. Our model correctly identified an average of 83.0% of the known beaver dams and 72.5% of the known beaver ponds based on validation with an independent dataset. This study demonstrates that remote sensing is an effective approach for identifying beaver-modified wetland features and can be applied to map these and other wetland habitat features of interest across large spatial extents. Furthermore, the systematic acquisition strategy of the remote sensors employed is well suited for monitoring changes in wetland conditions that affect the availability of habitats important to waterfowl and other wildlife.

Minimizing Data Waste: Conservation in the Big Data Era

Minimizing Data Waste: Conservation in the Big Data Era

Windmapper: An Efficient Wind Downscaling Method for Hydrological Models

AbstractEstimates of near‐surface wind speed and direction are key meteorological components for predicting many surface hydrometeorological processes that influence critical aspects of hydrological and biological systems. However, observations of near‐surface wind are typically spatially sparse. The use of these sparse wind fields to force distributed models, such as hydrological models, is greatly complicated in complex terrain, such as mountain headwaters basins. In these regions, wind flows are heavily impacted by overlapping influences of terrain at different scales. This can have a great impact on calculations of evapotranspiration, snowmelt, and blowing snow transport and sublimation. The use of high‐resolution atmospheric models allows for numerical weather prediction (NWP) model outputs to be dynamically downscaled. However, the computation burden for large spatial extents and long periods of time often precludes their use. Here, a wind‐library approach is presented to aid in downscaling NWP outputs and terrain‐correcting spatially interpolated observations. This approach preserves important spatial characteristics of the flow field at a fraction of the computational costs of even the simplest high‐resolution atmospheric models. This approach improves on previous implementations by: scaling to large spatial extents O(1M km2); approximating lee‐side effects; and fully automating the creation of the wind library. Overall, this approach was shown to have a third quartile RMSE of 1.8 and a third quartile RMSE of 58.2° versus a standalone diagnostic windflow model. The wind velocity estimates versus observations were better than existing empirical terrain‐based estimates and computational savings were approximately 100‐fold versus the diagnostic model.

Understanding the role of traditional and user-created recreation data in the cumulative footprint of recreation

Outdoor recreation is becoming more popular globally. However, recreation can contribute to biodiversity loss by modifying habitats and disturbing species. Understanding what activities are happening and where is crucial for mitigating negative impacts but is hampered by information gaps on patterns of human use, including the spatial footprint of recreation. Data on recreation trails and linear features traditionally have been managed in centralized government databases. Social media and user-created content, however, largely has been unexplored for tracking recreation patterns across large spatial extents, which is needed for landscape-level conservation (i.e., beyond a single protected area), planning, and management that meets the needs of people and wildlife. We compiled recreation data from government (documented) and non-government (undocumented) databases for motorized and non-motorized recreation in western Canada. Of all the trails mapped, 73% were classified as documented whereas 27% were undocumented. For undocumented trails, the primary data source was Open Street Maps, which is the basis for many recreation smartphone applications. Most undocumented trails had unknown activity types, indicating a lack of information in government databases about where recreation occurs, and which type(s) of activity occur there. Modeling revealed an increased probability of a trail being undocumented as elevation increased and the distance to nearest road decreased. Our results indicate that including data from user-created sources can improve estimates of the recreation human footprint. Management implications-Over 51% of trails and linear features were primarily used for motorized activities. This calls for better land-use planning to ensure high quality recreation experiences.-27% of our trail data originated from non-government data sources, primarily Open Street Maps. Undocumented trails were similarly likely to occur in protected areas (PA), as compared to unprotected areas. Better monitoring of the recreation footprint is required, inside and outside PAs, and data quality control methods implemented.-Likely, the situation will require a large-scale effort to centralize trail data, such as we have done here, strategically plan when and where recreation should occur, and educate user groups to mitigate ecological and wildlife effects.

Mapping Fire Severity in Southwest China Using the Combination of Sentinel 2 and GF Series Satellite Images.

Fire severity mapping can capture heterogeneous fire severity patterns over large spatial extents. Although numerous remote sensing approaches have been established, regional-scale fire severity mapping at fine spatial scales (<5 m) from high-resolution satellite images is challenging. The fire severity of a vast forest fire that occurred in Southwest China was mapped at 2 m spatial resolution by random forest models using Sentinel 2 and GF series remote sensing images. This study demonstrated that using the combination of Sentinel 2 and GF series satellite images showed some improvement (from 85% to 91%) in global classification accuracy compared to using only Sentinel 2 images. The classification accuracy of unburnt, moderate, and high severity classes was significantly higher (>85%) than the accuracy of low severity classes in both cases. Adding high-resolution GF series images to the training dataset reduced the probability of low severity being under-predicted and improved the accuracy of the low severity class from 54.55% to 72.73%. RdNBR was the most important feature, and the red edge bands of Sentinel 2 images had relatively high importance. Additional studies are needed to explore the sensitivity of different spatial scales satellite images for mapping fire severity at fine spatial scales across various ecosystems.

Hunting for Neighboring Open Clusters with Gaia DR3: 101 New Open Clusters within 500 pc

We systematically searched for open clusters in the solar neighborhood within 500 pc using the pyUPMASK and HDBSCAN clustering algorithms based on Gaia DR3. Taking into consideration that the physical size for most open clusters is less than 50 pc, we adopted a slicing approach for different distance shells and identified 324 neighboring open clusters, including 223 reported open clusters and 101 newly discovered open clusters, named here as the Open Cluster of Solar Neighborhood. Our discovery has increased the number of open clusters in the solar neighborhood by about 45%. In this work, larger spatial extents and more member stars were attained for our cluster sample. We provided the member stars and the membership probabilities through the pyUPMASK algorithm for each cluster and derived their astrophysical, age, and structural parameters.

Cross-boundary cooperation in wildfire management during the custodial management period of the US Forest Service: A case study of the eastern Cascades of Oregon, USA, 1905–1945

In the U.S., federal, tribal, state, local, and private land management entities seek to implement a wildfire management strategy that spans large spatial extents and multiple ownerships to achieve wildfire risk reduction and forest restoration. This strategy requires cross-boundary cooperation. Cooperative cross-boundary forest management is not new; yet little case study research has documented historical examples. This study provides insight into how potentially overlooked models of cross-boundary cooperation of the past may be applicable to renewed cross-boundary cooperation today. We examined cooperation during the custodial management period (1905–1945) of Forest Service history. This study identifies the governance institutions, partner roles, and practices that promoted the establishment and durability of cross-boundary cooperation under two governance models adopted by the Forest Service and partners. The models are: cooperative federalism, for partnering with state and private landowner local entities – forest protective associations (FPAs); and federal interagency cooperation, for partnering with the Bureau of Indian Affairs Forestry Division, historically responsible for managing tribal lands. Using case study methods, we examined two examples of each model from the eastern Cascades of Oregon, USA. Two governance characteristics promoted durable cooperation, regardless of model: (1) the development of a bundle of cooperation-focused institutions (cross-organizational policy alignment; headquarters-level cooperative policies, laws, and/ or interagency agreements; unit-level cooperative agreements) that linked multiple Forest Service administrative levels (agency headquarters, local unit) to those of partner entities; and (2) the performance of a diversity of partner roles— operating at both headquarters and local levels — that enabled management entities to offset capacity constraints. Several governance characteristics promoted cooperation in the model that engaged private landowners (cooperative federalism). These were: the state-granted authorities of a special purpose district, whereby self-organized FPAs garnered resources and developed and enforced cooperative practices that coordinated upward with state and federal policies; recognition by the state and Forest Service of authorized, local entities that represented private landowners (FPAs) as essential partners; an administrative structure for state transfer of federal funding to local entities, resulting in cost-sharing; and the state’s performance of the roles of cost-sharing administrator and liaison between Forest Service and local entities. We suggest that the more partners can integrate these features into cross-boundary wildfire management cooperation today, the better its chances of establishment and durability.

Structural diversity: a digital revolution

One of the fundamental rules in biology and ecology is that structure determines function. At a microscale, the transmissibility of COVID-19 is related to the unique structure of the novel coronavirus's trimeric spike protein. At a macroscale, the success of early human settlement along the west coast of North America during the Holocene is likely associated with the rich resources provided by the diverse structures of kelp forests. Until recently, research on ecosystem structure and diversity has lagged, due primarily to a lack of tools to measure these components effectively and accurately. For instance, scientists studying deciduous forests have relied on stem size and leaf area as proxies of forest structure to predict ecosystem functions. However, these traditional, proxy-based metrics are laborious to collect, logistically challenging to assess at large scales, and fail to capture other critical aspects of structural diversity such as canopy depth and complexity, both of which provide better predictions of ecosystem functions and processes as compared to their conventional counterparts. Fortunately, innovations in digital technology and data science are converging to yield a wide range of new research capabilities. As demonstrated by the papers in this Special Issue, ecologists are starting to embrace these promising advances, unlocking opportunities to understand structural diversity. We now have a robust suite of sensors that can be deployed on different platforms and used to map and measure three-dimensional (3D) structures in both terrestrial and aquatic systems, across a wide range of scales and resolutions. For example, modern single-photon lidar can provide wall-to-wall, high-resolution, 3D point cloud data in less than one hour for an area of over 10,000 ha, allowing individual tree-level measurement. Spanning large geographic regions, these unprecedented 3D data will enable investigations of fundamental ecological questions (eg macroscale patterns, scaling) that traditional plot-based approaches cannot. As with other innovations, however, these structural data also bring unencountered challenges. To take full advantage of ecology's digital advances, we argue that the following challenges must be addressed. The first is the necessity to ensure truly transdisciplinary collaboration. Although recognized for decades, the demand for such collaboration has never been greater in 3D structural diversity enabled ecology, especially at its current early stages of development. Most of the skillsets essential to harnessing and realizing the potential of these new technologies currently lie in disciplines beyond ecology and the environmental sciences. We need expertise from aviation technology for stable and safe data acquisition, engineering solutions for sensor integration, geomatics for data registration, computer science for data integration and segmentation, graphic design for visualization, information science for managing data and ensuring fair access, and social science for understanding societal implications and unintended consequences, to name a few. The quickest way for ecologists and environmental scientists to incorporate these research skills into their own work is through collaboration. The second challenge is embracing novel computational platforms (eg super computers, cloud computing) and approaches (eg machine learning and artificial intelligence). When covering large spatial extents, these 3D data often range from gigabytes to terabytes to petabytes in size. Therefore, advanced computing power and algorithms are required to handle them. Third, to better utilize these data, investigators will have to shift from traditional hypothesis-driven approaches to data-driven approaches or to a hybrid of the two. As these 3D data can generate hundreds of variables measuring different aspects of ecosystem structure, making sense of the data, especially when combined with other climatic and geophysical datasets, can be daunting or nearly impossible. Such massively multivariate datasets can quickly overwhelm traditional process-based modeling approaches. When deployed by knowledgeable analysts, machine learning approaches – whether data driven or domain knowledge guided – may reveal insights into ecosystem patterns and processes. Lastly, there is a pressing need to educate our next generation to embrace these digital advances. New data-oriented skills such as acquisition, visualization, analysis, and management of large datasets must become essential parts of ecological training. Consideration of structural diversity is foundational to our understanding of ecological systems. Consequently, advances in 3D data acquisition and analytics will likely reshape our understanding of ecology. Adopting these emerging digital technologies will enable the next generation of ecologists to gracefully operate a fleet of sensors to measure ecosystems and swim freely in the resulting ocean of data.