Canopy Tree Species Research Articles

Distinct latitudinal patterns of shifting spring phenology across the Appalachian Trail Corridor.

Warming associated with climate change will advance the onset of spring phenology for many forest plants across the Eastern United States. Understory forbs and spring ephemerals that fix a disproportionate amount of carbon during early spring may be negatively affected by earlier canopy closure; however, information on the spatial patterns of phenological change for these communities is still lacking. To assess the potential for changes in spring phenological windows, we synthesized observations from the Appalachian Mountain Club's (AMCs) Mountain Watch (MW) project, the National Phenology Network (NPN), and AMC's iNaturalist projects between 2004 and 2022 (n = 118,250) across the length of the Appalachian Trail (AT) Corridor (34° N-46° N latitude). We used hierarchical Bayesian modeling to examine the sensitivity of spring flowering and leaf-out for 11 understory species and 14 canopy tree species to mean spring temperature (April-June). We conducted analyses across the AT Corridor, partitioned by regions of 4° latitude (south, mid-Atlantic, and north). Spring phenologies for both understory plants and canopy trees advanced with warming (~6 and ~3 days/°C, respectively). However, the sensitivity of each group varied by latitude, with the phenology of trees and understory plants advancing to a greater degree in the mid-Atlantic region (~10 days/°C) than in the southern or northern regions (~5 days/°C). While we find evidence that phenological windows remain stable in the southern and mid-Atlantic portions of the AT, we observed an expansion of the spring phenological window in the north where there was greater understory forb temperature sensitivity compared with trees (~2.7 days/°C). Our analyses indicate the differential sensitivity of forest plant phenology to potential warming across a large latitudinal gradient in the Eastern United States. Further, evidence for a temperature-driven expansion of the spring phenological window suggests a potential beneficial effect for understory plants in the northern AT, although phenological mismatch with potential pollinators and increased vulnerability to late winter frosts are possible. Using extensive citizen-science datasets allows us to synthesize regional- and continental-scale data to explore spatial and temporal trends in spring phenology related to warming. Such data can help to standardize approaches in phenological research and its application to forest climate resiliency.

The space occupation and use by tree crowns explain variations of individual growth rates in an old-growth temperate forest in Japan

Relative growth rates (RGR), i.e., growth rates per unit biomass (ΔM/M, where M is plant mass and ΔM is the change in M over a period of time), reflect growth strategies of plant species. Partitioning of RGR to net assimilation rate (ΔM/leaf area) and leaf area ratio (leaf area/M) provides further insights into allocation strategies. To apply this analytical approach to large canopy tree species, we used crown area (Ac) as a proxy for leaf area to understand variations of RGR partitioned to space use efficiency (SUE, ΔM/Ac) and space occupation efficiency (SOE, Ac/M). With UAV imagery, we measured Ac of 226 co-occurring individuals of 14 canopy tree species in a 1-ha stand in a temperate old-growth mixed-forest in Japan, and analyzed how RGR was related to SUE and SOE. The results show that deciduous species exhibited higher SOE and lower SUE compared to evergreen species, even though their RGR values largely overlapped. Late successional species tended to have higher RGR through higher SUE than early-to-middle successional species. We also analyzed the relationship of absolute growth rates (AGR) with several functional traits including DBH (diameter at breast height), Ac, leaf- and stem traits. Both Ac and DBH were strong determinants of AGR across species. Low specific leaf area (SLA, leaf area per unit leaf mass) and high wood density positively contributed to AGR across species, offering long-term growth advantages for old-growth natural forest. The analytical framework introduced in this study may be useful to elucidate the variation of tree growth strategies of canopy trees in natural forest stands.

Large-scale changes in the distribution of suitable habitat of the endangered subtropical canopy tree species Vatica guangxiensis under climate change

Large-scale changes in the distribution of suitable habitat of the endangered subtropical canopy tree species Vatica guangxiensis under climate change

Influence of Native Woody Understory on Invasive Grasses and Soil Nitrogen Dynamics Under Plantation and Remnant Montane Tropical Trees

While the influence of canopy trees on soils in natural and restored forest environments is well studied, the influence of understory species is not. Here, we evaluate the effects of outplanted native woody understory on invasive grass biomass and soil nutrient properties in heavily grass-invaded 30 + year-old plantations of a native N-fixing tree Acacia koa in Hawai‘i. We analyze soils from under A. koa trees with versus without planted woody understory and compare these to soils from under remnant pasture trees of the pre-deforestation dominant, Metrosideros polymorpha where passive recruitment of native woody understory has occurred since the cessation of grazing. Simultaneously, we experimentally planted understory species at three times the density used by managers to see if this could quickly decrease grass biomass and change soil nutrient dynamics. We found that invasive grass biomass declined with understory planting in surveyed and experimental sites. Yet, woody understory abundance had no effect on N cycling. Short-term N availability and nitrification potential were higher under A. koa than M. polymorpha trees regardless of understory. Net N mineralization either did not differ (~ 1 mo) between canopy species or was higher (171 day incubations) under remnant M. polymorpha where organic matter was also higher. The only influence of understory on soil was a positive correlation with loss-on-ignition (organic matter) under M. polymorpha. We also demonstrate differential controls over N cycling under the two canopy tree species. Overall, understory restoration has not changed soil characteristics even as invasive grass biomass declines.

Individual canopy tree species maps for the National Ecological Observatory Network.

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 km2 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 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.

Understory plants evade shading in a temperate deciduous forest amid climate variability by shifting phenology in synchrony with canopy trees.

Global warming is leading understory and canopy plant communities of temperate deciduous forests to grow leaves earlier in spring and drop them later in autumn. If understory species extend their leafy seasons less than canopy trees, they will intercept less light. We look for mismatched phenological shifts between canopy and understory in 28 years (1995-2022) of weekly data from Trelease Woods, Urbana, IL, USA. The observations cover 31 herb species of contrasting seasonality (for 1995-2017), three sapling species, and the 15 most dominant canopy tree species for all years, combined with solar radiation, temperature and canopy light transmittance data. We estimate how understory phenology, cold temperatures, canopy phenology, and solar radiation have individually limited understory plants' potential light interception over >2 decades. Understory and canopy phenology were the two factors most limiting to understory light availability, but which was more limiting varied greatly among species and among/within seasonality groups; solar radiation ranked third and cold fourth. Understory and canopy phenology shifts usually occurred in the same direction; either both strata were early or both were late, offsetting each other's effects. The four light-limiting factors combined showed significant temporal trends for six understory species, five toward less light interception. Warmer springs were significantly associated with shifts toward more light interception in three sapling species and 19 herb species. Canopy phenology became more limiting in warmer years for all three saplings species and 31 herb species. However, in aggregate, these variables mostly offset one another; only one sapling and seven herb species showed overall significant (and negative) relationships between light interception and spring temperature. The few understory species mismatched with canopy phenology due to changing climate are likely to intercept less light in future warmer years. The few species with data for carbon assimilation show broadly similar patterns to light interception.

Forest practitioners’ requirements for remote sensing-based canopy height, wood-volume, tree species, and disturbance products

Abstract Despite decades of development, the uptake of remote sensing-based information products in the forestry sector is still lagging behind in central and southern Europe. This may partly relate to a mismatch of the developed remote sensing products and the requirements of potential users. Here, we present the results of a questionnaire survey in which we questioned 355 forest practitioners from eight central and southern European countries. We aimed to learn about forest practitioners' technical requirements for four remote sensing-based information products, including information on tree species, canopy height, wood volume/biomass, and forest disturbances. We asked for practitioners’ preferences with respect to thematic and spatial detail as well as the maximal acceptable error and the temporal frequency with which the information layers would be needed. We then examined whether the education, age, and professional background affect the requirements. Preferences with respect to spatial and thematic detail were comparably diverse while more homogenous patterns could be observed for demands with respect to errors and temporal frequency. Our results indicate that for some information products such as canopy height maps, existing remote sensing technology, and workflows can match all demands of practitioners. Remotely sensed information on forest disturbances partly fulfils the demands of the practitioners while for products related to tree species and wood volume/biomass the level of thematic detail and the accuracy of the products demanded by practitioners in central and southern Europe is not yet fully matched. We found no statistically significant differences between the demographic groups examined. The findings of this study improve our understanding of matches and mismatches of the technical requirements of practitioners for remote sensing-based information products.

Diel activity patterns of a canopy-inhibiting beetle community (Coleoptera) in a Neotropical rainforest

Diel activity is one main feature of animal‘s behavior and is often an intrinsic trait characterizing distinct taxonomic groups. Abiotic conditions such as temperature may influence the diel activity patterns of arthropod communities associated with a particular ecosystem or habitat. Similarly, biotic factors, such as resource availability, affect arthropod activity. In addition, diel activity is thought to be an important factor in niche partitioning of arthropod communities. As part of a larger beetle survey in a lowland tropical rainforest in southern Venezuela, I analyzed the diel activity of an arboreal beetle community collected from 23 canopy-tree species over a cumulative year. Diel activity was observed in 535 beetle species, comprising 5,948 individuals, using a canopy crane installed in the study area. Of the 535 beetle species, 198 (37%) showed diurnal activity, and 281 (52.5%) showed nocturnal activity. In contrast, the proportions of nocturnal (n = 2,024, 34%) and diurnal (n = 1,983, 33.3%) individuals were balanced. Most of the observed beetles occurred only during the activity phase in their host trees. This particularly applies to extrafloral nectary- and flower-visiting beetle species. Flowering trees attracted different proportions of diurnal and nocturnal species according to flowering syndrome, whereas extrafloral nectaries were mainly visited at night. Thus, the beetle communities associated with single tree species showed distinct compositions of nocturnal and diurnal species.

Seedling dynamics differ between canopy species and understory species in a tropical seasonal rainforest, SW China

We used 11 years of census data from 450 seedling quadrats established in a 20-ha forest dynamics plot to study seedling dynamics in tree species of a tropical seasonal rainforest in Xishuangbanna, southwestern China. We found that overall seedling recruitment rate and relative growth rate were higher in the rainy season than in the dry season. Both the recruitment rate of seedlings from canopy tree species (two species) and the relative growth rate of seedlings from understory species (nine species) were higher in the rainy season than in the dry season. However, in the rainy season, the recruitment rate of seedlings was higher for canopy tree species than for understory tree species. In addition, relative growth rate of seedlings was higher in the canopy species than in understory seedlings in the dry season. We also observed that, in both rainy and dry seasons, mortality rate of seedlings was higher for canopy species than for understory species. Overall, canopy tree species appear to have evolved a flexible strategy to adapt to the seasonal changes of a monsoon climate. In contrast, understory tree species seem to have adopted a conservative strategy. Specifically, these species mainly release seedlings in the rainy season and maintain relatively stable populations with a lower mortality rate and recruitment rate in both dry and rainy seasons. Our study suggests that canopy and understory seedling populations growing in forest understory may respond to future climate change scenarios with distinct regeneration strategies.

Generalist southern African temperate forest canopy tree species have distinct pollinator communities partially predicted by floral traits

AbstractForest canopies provide important resources for insect communities via flowers. Yet, pollination systems of tall forest trees are poorly studied, resulting from the difficulties in observing pollinator activity at the canopy level and great temporal variation in flower production. In temperate forest canopies of the southern hemisphere, small, whitish and generalist flowers seem to dominate. Here, we observed insect flower visitors, at the canopy level, to four southern Afrotemperate forest tree species bearing small, white to green flowers in a large, indigenous forest. Additionally, we quantified flower traits and collected pollen from representative insect visitors. A total of 105 insect species, from 48 families and 7 orders, were observed visiting flowers. In terms of total flower visits, the generalist Cape honey bee (Apis mellifera capensis) made up ca. 57% of all flower visits. A third of the total observation time covered crepuscular to nocturnal flower visits; yet only 12.68% of total visits took place during this time. Interestingly, despite both trees and insects being largely generalist in their interactions with one another (supported by the presence of conspecific and heterospecific pollen on most flower visitors), some insect species showed strong preferences for specific species of tree, driving dissimilar, interspecific assemblages of flower visitors. The pollinator community disparity may be explained through the unique and dissimilar floral traits for each tree species, both in flower size and in petal reflectance. We conclude that within generalist pollination systems, distinct and non‐random mutualisms can develop between different species of plants and a diverse suite of pollinators, and that floral traits could partially predict such interactions.

Natural shading vs. artificial shading: A comparative analysis of their cooling efficacy in extreme hot weather

Shading, comprising both natural and artificial coverings, is an efficient and effective strategy to reduce heat stress in outdoor environments. Despite its recognized efficiency, there is limited comparative insight into the cooling performance of natural and artificial shading. This study examined the cooling efficacy of two covered walkways and two tree canopies, concerning microclimate conditions by various variables and thermal stress quantified by four thermal indices. Our findings highlight three key points: Firstly, the average cooling effects provided by natural and artificial shading were comparable, including reductions in air temperature (1.42 °C vs. 1.31 °C), mean radiant temperature (15.93 °C vs. 13.71 °C), and Physiological Equivalent Temperature (PET) (9.06 °C vs. 9.70 °C), Universal Thermal Climate Index (UTCI) (4.71 °C vs. 5.08 °C), and Hong Kong Heat Index (HKHI) (2.36 °C vs. 2.50 °C). Secondly, it is inconsistent which shading solution is better for microclimate improvement and heat stress alleviation, which is affected by canopy materials, configurations, and tree species. Thirdly, the four indices presented an inconsistent pattern in heat stress assessment. Notably, PET and UTCI showed similar patterns, while HKHI and modified PET presented lower sensitivity to high heat stress. This study offers valuable insights for urban designers to create sustainable shaded communities and also initiates a discourse for revising local heat warning systems, considering various thermal stress indices.

Seasonal and long-term climate drivers of tree species phenology and litterfall in a Nothofagus cool temperate rainforest of Australia

The cool temperate rainforests of eastern Australia are at risk from anthropogenic climate change with predicted changes in temperature, rainfall, severe weather, basal cloud layer, and droughts. Phenology and litter production are fundamental reproductive and growth processes to document in any ecosystem, yet very few long-term studies exist in Australian rainforests. In this study, long-term datasets are used to describe different phenological and litter production behaviours of tree species in a Nothofagus cool temperate rainforest in New South Wales (NSW), Australia, analysing seasonal and inter-annual climate drivers. Leaf fall at the community level was mostly influenced by Nothofagus moorei, driven by temperature and wind speed, and Ceratopetalum apetalum, driven by temperature, rainfall, and solar radiation. Mean dates of leaf fall at the community level were found to be advancing, correlated with an advance in solar radiation. We also analysed in detail the flowering behaviour of the dominant canopy tree species, N. moorei, which masts with a mean inter-flowering period of 3–4 years in 65% of flowering events. Three of the studied species presented mast flowering, C. apetalum, N. moorei, and Orites excelsus; however, they did not mast in the same years. All species presented strong seasonality in their phenological activity, but seasonality peaked in different months, and were driven by varied climate variables. Supra-annual peaks of flowering and fruiting did not occur at the same time for all species, and climate drivers of inter-annual phenological behaviour were different for each species. Our results show that projected changes in climate will affect species from cool temperate rainforests differently, affecting not only biomass production, but also species reproductive output and forest dynamics.

How tree stand phenology determines understorey senescence - a case study from boreal forests

Leaf fall in the autumn opens the forest canopy, allowing more solar radiation to be transmitted to the forest floor. Those understorey species that remain physiologically active into the autumn may benefit from the sunlight received by extending their growing season, to assimilate additional carbon while conditions remain favourable. We monitored leaf water and pigment content, as well as photosynthetic capacity in understorey species growing in adjacent stands differing in their canopy tree species. Leaf fall, transmitted light, and microclimate were monitored in each stand. We found that overstorey leaf fall started earlier in the birch (Betula pendula, L.), than in the oak (Quercus robur, L.) stand, and light transmission changed accordingly. Concurrently, understorey leaf senescence was generally earlier in the birch than in the oak stand, itself earlier than in the evergreen spruce stand (Picea abies, L. H. Karst.). Neither atmospheric CO2, humidity, nor temperature differed between stands. A change in light quality and/or increase in quantity following leaf fall drove the difference in the timing of senescence in the understorey. Understorey species with later senescence were able to use the increased light more after leaf fall. Together these findings help to provide a mechanistic foundation to predict how ecosystem functioning and ultimately carbon balance will be impacted by phenological shifts in response to global changes.

Age-Dating of Foliage and Soil Organic Matter: Aligning 228Th:228Ra and 7Be:210Pb Radionuclide Chronometers over Annual to Decadal Time Scales.

The 228Th:228Ra ratios of foliage and organic soil horizons evolve with time following a predictable radioactive decay law and thus provide a new chronometer for absolute age-dating of plant and soil organic matter. Preferential uptake of 228Th (t0.5 = 1.9 years) and 228Ra (t0.5 = 5.9 years) by canopy tree species, ferns, and mosses, drives disequilibrium in the 232Th-228Ra-228Th radioactive decay series within forest vegetation and organic soils. With examples from northeastern USA, we verify a new 228Th:228Ra age model by demonstrating its concordance with the fallout radionuclide chronometer 7Be:210Pb in the 0 to 5-year time frame [R2 = 0.87, RMSE = 0.5 years]. At our locality, canopy tree species assimilate 228Th with a typical initial ratio (228Th:228Ra)0 ∼ 0.3, but in several instances, both deciduous and coniferous tree species show a preference for Th over Ra with (228Th:228Ra)0 exceeding 5. While the 228Th:228Ra system is restricted to organic soil horizons, concordance of 228Th:228Ra with established 7Be:210Pb and 241Am bomb-pulse chronometers establishes a coherent age-dating system of soil organic matter based on three independent chronometers and five particle reactive metals, and spanning 0-200 years in time scale that encompasses both organic and mineral soils to depths of up to 30 cm. Concordance indicates that these metals all follow common processes of organometallic colloid formation and migration and, in conjunction with 14C, may open new opportunities to understand soil pedogenic processes that regulate the storage of carbon and atmospheric metals such as Pb and Hg.

Altitudinal Patterns of Foliar Nutrients in Dominant Tree Species: Are Central Himalayan Forests Nutrient Stressed?

In forest ecosystems nutrient availability is one of the most important drivers of tree growth and ecosystem function. We studied changes in soil and foliar nutrients (N, P & K) and leaf N and P stoichiometry in eight dominant canopy and sub-canopy forest tree species of Central Himalayan forests occurring between 300 and 2200 m asl. Nutrient (N, P & K) concentration in leaves at bud break, mature and senescent stage varied significantly across the tree species. The nutrient resorption efficiency (NRE) varied significantly across the species and found ranging for N from 48.7 - 76.8%. Both the nitrogen resorption efficiency (NRE) and phosphorus resorption efficiency (PRE) increased linearly with altitude (R2= 0.647 for NRE significant at p< 0.05; and R2 for PRE= 0.525; NS) pointing out that trees growing in temperate climate conserve nutrients more efficiently to withstand the slow mineralization due to cold climate. With the increasing altitude leaf life-span increases significantly (R2= 0.0025; p< 0.0074). ANOVA indicates that the foliar N and P and altitude were unrelated. The low N: P ratio, a measure of nutrient status of forest ecosystems at senescent leaf stage, found ranging from 9.97 in Q. floribunda to 18.8 in P. roxburghii points out that forests of the study region are nutrient stressed, partly due to low P mineralization in temperate climate, leaching of nutrients due to heavy monsoon rain, and poorly developed underlying geology. Thus, the Central Himalayan forests are characterized by high NRE and low growth rate with increasing altitude. Winter season warming might help improve the nutrient availability in the forest floor and nutrient uptake to enhance nutrient cycling and biomass productivity in these forests.

Comparative physiology of canopy tree leaves in evergreen and deciduous forests in lowland Thailand

The typical seasonally dry forests in Southeast Asia are the mixed deciduous forest (MDF), dry dipterocarp (deciduous) forest (DDF), and dry evergreen forest (DEF). We obtained 21 physiological traits in the top/sunlit leaves of 107, 65 and 51 tree species in MDF, DEF and DDF, respectively. Approximately 70%, 95% and 95% of canopy tree species which consist of MDF, DEF and DDF are sampled, respectively. Light-saturated photosynthetic rates (Asat) exhibit a positive correlation with foliar nitrogen (N) and phosphorus (P) on leaf mass and area bases across tree species. Decreased leaf mass-based P reduces the positive slope of the mass-based N and Asat relationship across species and habitats. The differences in nutrient and water use and leaf habits are well matched to the variation in soil properties among the forest types, highlighting the reliability of this comprehensive database for revealing the mechanism of niche segregation based on edaphic factors.

Multi‐scale habitat overlap in two broad‐ranged sympatric Neotropical forest eagles reveals shared environmental space and habitat use

Quantifying resource partitioning between co‐occurring species can provide insight into processes facilitating coexistence by closely related species, a fundamental question in ecology. We tested whether the habitat requirements of two closely related Neotropical forest eagles, the Crested Eagle Morphnus guianensis and Harpy Eagle Harpia harpyja, differ at fine and coarse resolutions across their shared geographical range. Using landcover and topographical covariates, we quantified potential resource overlap first using higher resolution (30 arc‐s, ~ 1‐km2 data) generalized linear models (GLMs), and secondly using coarser‐grain (2.5 arc‐min, ~ 4.5‐km2 data) environmental ordination to capture the potential effect of scale on habitat overlap. The distribution of both eagles was largely explained by canopy tree species richness and canopy structural complexity, with peak suitability of 60–80% evergreen forest cover. Both eagles were negatively associated with mosaic forest and cultivated areas. From the GLMs, habitat overlap was >93% in geographical space but was reduced to 73% when considering environmental space, a proxy for resource overlap. From ordination (principal component analysis), resource overlap was 67% in environmental space, with randomization tests supporting equivalent environmental space for both eagles. Our results suggest that at the continental scale, Crested and Harpy Eagles share identical environmental space when quantified at fine and broad scales, with little difference in distribution and habitat use. At the continental scale used here, both eagles can coexist, presumably with sufficient habitat heterogeneity for coexistence when they occur in close proximity. Therefore, further research is required at the local level to capture fully where coexistence at the local scale is facilitated more by fine‐scale habitat selection, or difference in diet between two species with indistinguishable habitat use.

Urban green and blue infrastructure effect on the micro-scale thermal environment in a residential neighborhood: Mueller, Austin, TX

ABSTRACT This study examined the role and importance of various green and blue infrastructure elements in improving the deteriorating urban thermal environment from a post-occupancy evaluation perspective of residential neighborhood development. The results showed that the influential variables in this study were the tree canopy coverage ratio (TCR), tree species, yard area, and distance to a waterbody. In addition, a house with a large yard area increases the amount of direct exposure to solar radiation, so the comfort of the outdoor thermal environment was low. Furthermore, the distance to a waterbody parameter did not have a positive impact on the outdoor thermal comfort level because the dominant wind direction within the study area during the day was opposite to the blue infrastructure. Lastly, geographically weighted regression (GWR) analysis was adopted to identify an improvement to compensate for the relatively low explanatory nature of the ordinary least squares (OLS) models due to spatial correlation.

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.

Capturing long‐tailed individual tree diversity using an airborne imaging and a multi‐temporal hierarchical model

AbstractMeasuring forest biodiversity using terrestrial surveys is expensive and can only capture common species abundance in large heterogeneous landscapes. In contrast, combining airborne imagery with computer vision can generate individual tree data at the scales of hundreds of thousands of trees. To train computer vision models, ground‐based species labels are combined with airborne reflectance data. Due to the difficulty of finding rare species in a large landscape, many classification models only include the most abundant species, leading to biased predictions at broad scales. For example, if only common species are used to train the model, this assumes that these samples are representative across the entire landscape. Extending classification models to include rare species requires targeted data collection and algorithmic improvements to overcome large data imbalances between dominant and rare taxa. We use a targeted sampling workflow to the Ordway Swisher Biological Station within the US National Ecological Observatory Network (NEON), where traditional forestry plots had identified six canopy tree species with more than 10 individuals at the site. Combining iterative model development with rare species sampling, we extend a training dataset to include 14 species. Using a multi‐temporal hierarchical model, we demonstrate the ability to include species predicted at <1% frequency in landscape without losing performance on the dominant species. The final model has over 75% accuracy for 14 species with improved rare species classification compared to 61% accuracy of a baseline deep learning model. After filtering out dead trees, we generate landscape species maps of individual crowns for over 670 000 individual trees. We find distinct patches of forest composed of rarer species at the full‐site scale, highlighting the importance of capturing species diversity in training data. We estimate the relative abundance of 14 species within the landscape and provide three measures of uncertainty to generate a range of counts for each species. For example, we estimate that the dominant species, Pinus palustris accounts for c. 28% of predicted stems, with models predicting a range of counts between 160 000 and 210 000 individuals. These maps provide the first estimates of canopy tree diversity within a NEON site to include rare species and provide a blueprint for capturing tree diversity using airborne computer vision at broad scales.