Search for: All records

These data were collected to assess how seed availability and site limitations affect conifer germination across species distributions. Our study focused on areas above alpine treeline where subalpine tree species must migrate to track movement of suitable climate, but we also included sites in the core and at the lower ecotone of subalpine forests. We monitored seed availability and germination of new seedlings for four subalpine tree species from 2015-present at Niwot Ridge, Colorado, USA. Seed availability was collected in 66-95 seed traps in 14-17 sites (6-12 traps per site; see data for count per site), depending on year. In the lab, seeds were counted by species. In the field, new germinants were counted by species 3-5 weeks after snow disappearance (i.e., peak germination) and again in late September from 2015 to 2018 only. Only one census of new germinants was conducted from 2019 to 2023. New germinants from prior years were censused in subsequent summers. 

The ecology of forest ecosystems depends on the composition of trees. Capturing fine-grained information on individual trees at broad scales provides a unique perspective on forest ecosystems, forest restoration, and responses to disturbance. Individual tree data at wide extents promises to increase the scale of forest analysis, biogeographic research, and ecosystem monitoring without losing details on individual species composition and abundance. Computer vision using deep neural networks can convert raw sensor data into predictions of individual canopy tree species through labeled data collected by field researchers. Using over 40,000 individual tree stems as training data, we create landscape-level species predictions for over 100 million individual trees across 24 sites in the National Ecological Observatory Network (NEON). Using hierarchical multi-temporal models fine-tuned for each geographic area, we produce open-source data available as 1 km²shapefiles with individual tree species prediction, as well as crown location, crown area, and height of 81 canopy tree species. Site-specific models had an average performance of 79% accuracy covering an average of 6 species per site, ranging from 3 to 15 species per site. All predictions are openly archived and have been uploaded to Google Earth Engine to benefit the ecology community and overlay with other remote sensing assets. We outline the potential utility and limitations of these data in ecology and computer vision research, as well as strategies for improving predictions using targeted data sampling. 

# Individual Tree Predictions for 100 million trees in the National Ecological Observatory Network Preprint: https://www.biorxiv.org/content/10.1101/2023.10.25.563626v1 ## Manuscript Abstract The ecology of forest ecosystems depends on the composition of trees. Capturing fine-grained information on individual trees at broad scales allows an unprecedented view of forest ecosystems, forest restoration and responses to disturbance. To create detailed maps of tree species, airborne remote sensing can cover areas containing millions of trees at high spatial resolution. Individual tree data at wide extents promises to increase the scale of forest analysis, biogeographic research, and ecosystem monitoring without losing details on individual species composition and abundance. Computer vision using deep neural networks can convert raw sensor data into predictions of individual tree species using ground truthed data collected by field researchers. Using over 40,000 individual tree stems as training data, we create landscape-level species predictions for over 100 million individual trees for 24 sites in the National Ecological Observatory Network. Using hierarchical multi-temporal models fine-tuned for each geographic area, we produce open-source data available as 1km^2 shapefiles with individual tree species prediction, as well as crown location, crown area and height of 81 canopy tree species. Site-specific models had an average performance of 79% accuracy covering an average of six species per site, ranging from 3 to 15 species. All predictions were uploaded to Google Earth Engine to benefit the ecology community and overlay with other remote sensing assets. These data can be used to study forest macro-ecology, functional ecology, and responses to anthropogenic change. ## Data Summary Each NEON site is a single zip archive with tree predictions for all available data. For site abbreviations see: https://www.neonscience.org/field-sites/explore-field-sites. For each site, there is a .zip and .csv. The .zip is a set 1km .shp tiles. The .csv is all trees in a single file. ## Prediction metadata *Geometry* A four pointed bounding box location in utm coordinates. *indiv_id* A unique crown identifier that combines the year, site and geoindex of the NEON airborne tile (e.g. 732000_4707000) is the utm coordinate of the top left of the tile.  *sci_name* The full latin name of predicted species aligned with NEON's taxonomic nomenclature.  *ens_score* The confidence score of the species prediction. This score is the output of the multi-temporal model for the ensemble hierarchical model.  *bleaf_taxa* Highest predicted category for the broadleaf submodel *bleaf_score* The confidence score for the broadleaf taxa submodel  *oak_taxa* Highest predicted category for the oak model  *dead_label* A two class alive/dead classification based on the RGB data. 0=Alive/1=Dead. *dead_score* The confidence score of the Alive/Dead prediction.  *site_id* The four letter code for the NEON site. See https://www.neonscience.org/field-sites/explore-field-sites for site locations. *conif_taxa* Highest predicted category for the conifer model *conif_score* The confidence score for the conifer taxa submodel *dom_taxa* Highest predicted category for the dominant taxa mode submodel *dom_score* The confidence score for the dominant taxa submodel ## Training data The crops.zip contains pre-cropped files. 369 band hyperspectral files are numpy arrays. RGB crops are .tif files. Naming format is __, for example. "NEON.PLA.D07.GRSM.00583_2022_RGB.tif" is RGB crop of the predicted crown of NEON data from Great Smoky Mountain National Park (GRSM), flown in 2022.Along with the crops are .csv files for various train-test split experiments for the manuscript. ### Crop metadata There are 30,042 individuals in the annotations.csv file. We keep all data, but we recommend a filtering step of atleast 20 records per species to reduce chance of taxonomic or data cleaning errors. This leaves 132 species. *score* This was the DeepForest crown score for the crop. *taxonID*For letter species code, see NEON plant taxonomy for scientific name: https://data.neonscience.org/taxonomic-lists *individual*unique individual identifier for a given field record and crown crop *siteID*The four letter code for the NEON site. See https://www.neonscience.org/field-sites/explore-field-sites for site locations. *plotID* NEON plot ID within the site. For more information on NEON sampling see: https://www.neonscience.org/data-samples/data-collection/observational-sampling/site-level-sampling-design *CHM_height* The LiDAR derived height for the field sampling point. *image_path* Relative pathname for the hyperspectral array, can be read by numpy.load -> format of 369 bands * Height * Weight *tile_year*  Flight year of the sensor data *RGB_image_path* Relative pathname for the RGB array, can be read by rasterio.open() # Code repository The predictions were made using the DeepTreeAttention repo: https://github.com/weecology/DeepTreeAttentionKey files include model definition for a [single year model](https://github.com/weecology/DeepTreeAttention/blob/main/src/models/Hang2020.py) and [Data preprocessing](https://github.com/weecology/DeepTreeAttention/blob/cae13f1e4271b5386e2379068f8239de3033ec40/src/utils.py#L59). 

Changes in climate are altering disturbance regimes in forests of western North America, leading to increases in the potential for disturbance events to overlap in time and space. Though interactions between abiotic and biotic disturbance (e.g., the effect of bark beetle outbreak on subsequent wildfire) have been widely studied, interactions between multiple biotic disturbances are poorly understood. Defoliating insects, such as the western spruce budworm (WSB; Choristoneura freemanni), have been widely suggested to predispose trees to secondary colonization by bark beetles, such as the Douglas-fir beetle (DFB; Dendroctonus pseudotsugae). However, there is little quantitative research that supports this observation. Here, we asked: Does previous WSB damage increase the likelihood of subsequent DFB outbreak in Douglas-fir (Pseudotsuga menziesii) forests of the Southern Rocky Mountains, USA? To quantify areas affected by WSB and then DFB, we analyzed Aerial Detection Survey data from 1999–2019. We found that a DFB presence followed WSB defoliation more often than expected under a null model (i.e., random distribution). With climate change expected to intensify some biotic disturbances, an understanding of the interactions between insect outbreaks is important for forest management planning, as well as for improving our understanding of forest change. 

Since the late 1990s, extensive outbreaks of native bark beetles (Curculionidae: Scolytinae) have affected coniferous forests throughout Europe and North America, driving changes in carbon storage, wildlife habitat, nutrient cycling, and water resource provisioning. Remote sensing is a crucial tool for quantifying the effects of these disturbances across broad landscapes. In particular, Landsat time series (LTS) are increasingly used to characterize outbreak dynamics, including the presence and severity of bark beetle-caused tree mortality, though broad-scale LTS-based maps are rarely informed by detailed field validation. Here we used spatial and temporal information from LTS products, in combination with extensive field data and Random Forest (RF) models, to develop 30-m maps of the presence (i.e., any occurrence) and severity (i.e., cumulative percent basal area mortality) of beetle-caused tree mortality 1997–2019 in subalpine forests throughout the Southern Rocky Mountains, USA. Using resultant maps, we also quantified spatial patterns of cumulative tree mortality throughout the region, an important yet poorly understood concept in beetle-affected forests. RF models using LTS products to predict presence and severity performed well, with 80.3% correctly classified (Kappa = 0.61) and R2 = 0.68 (RMSE = 17.3), respectively. We found that ≥10,256 km2 of subalpine forest area (39.5% of the study area) was affected by bark beetles and 19.3% of the study area experienced ≥70% tree mortality over the twenty-three year period. Variograms indicated that severity was autocorrelated at scales < 250 km. Interestingly, cumulative patch-size distributions showed that areas with a near-total loss of the overstory canopy (i.e., ≥90% mortality) were relatively small (<0.24 km2) and isolated throughout the study area. Our findings help to inform an understanding of the variable effects of bark beetle outbreaks across complex forested regions and provide insight into patterns of disturbance legacies, landscape connectivity, and susceptibility to future disturbance. 

1. Amplified by warming temperatures and drought, recent outbreaks of native bark beetles (Curculionidae: Scolytinae) have caused extensive tree mortality throughout Europe and North America. Despite their ubiquitous nature and important effects on ecosystems, forest recovery following such disturbances is poorly understood, particularly across regions with varying abiotic conditions and outbreak effects. 2. To better understand post-outbreak recovery across a topographically complex region, we synthesized data from 16 field studies spanning subalpine forests in the Southern Rocky Mountains, USA. From 1997 to 2019, these forests were heavily affected by outbreaks of three native bark beetle species (Dendroctonus ponderosae, Dendroctonus rufipennis and Dryocoetes confusus). We compared pre- and post-outbreak forest conditions and developed region-wide predictive maps of post-outbreak (1) live basal areas, (2) juvenile densities and (3) height growth rates for the most abundant tree species – aspen (Populus tremuloides), Engelmann spruce (Picea engelmannii), lodgepole pine (Pinus contorta) and subalpine fir (Abies lasiocarpa). 3. Beetle-caused tree mortality reduced the average diameter of live trees by 28.4% (5.6 cm), and species dominance was altered on 27.8% of field plots with shifts away from pine and spruce. However, most plots (82.1%) were likely to recover towards pre-outbreak tree densities without additional regeneration. Region-wide maps indicated that fir and aspen, non-host species for bark beetle species with the most severe effects (i.e. Dendroctonus spp.), will benefit from outbreaks through increased compositional dominance. After accounting for individual size, height growth for all conifer species was more rapid in sites with low winter precipitation, high winter temperatures and severe outbreaks. 4. Synthesis. In subalpine forests of the US Rocky Mountains, recent bark beetle outbreaks have reduced tree size and altered species composition. While eventual recovery of the pre-outbreak forest structure is likely in most places, changes in species composition may persist for decades. Still, forest communities following bark beetle outbreaks are widely variable due to differences in pre-outbreak conditions, outbreak severity and abiotic gradients. This regional variability has critical implications for ecosystem services and susceptibility to future disturbances.