Philippa Kaur

Knowing the distribution of species is crucial for effective conservation action. However, accurate and high-resolution spatial data are only available for a limited number of species1,2. For mammals and birds, the most comprehensive and widely used global distribution dataset is the set of range maps compiled as part of the assessments for the International Union for Conservation of Nature (IUCN) Red List. These represent each species’ distributional limits and tend to minimize omission errors (i.e. false absences) at the expense of commission errors (i.e. false presences)3,4. Therefore, they often contain sizeable areas not regularly occupied by the species.Maps of the Area of Habitat (AOH; previously known as Extent of Suitable Habitat, ESH) complement range maps by indicating potential occupancy within the range, thereby reducing commission errors5. AOH is defined as ‘the habitat available to a species, that is, habitat within its range’5. These models are produced by subtracting areas unsuitable for the species within their range, using information on each species’ associations with habitat and elevation5,6,7,8. Comprehensive sets of AOH maps have been produced in the past for mammals6 and amphibians7, as well as subsets of birds8,9. The percentage of a species’ range covered by the AOH varies depending on the methodology used to associate species to their habitats, and their habitats to land-cover, the coarseness of the range map, the region in which the species is distributed, and the species’ habitat specialization and elevation limits5. For example, Rondinini et al.6 found that, when considering elevation and land cover features for terrestrial mammals, the AOH comprised, on average, 55% of the range. Ficetola et al.7 obtained a similar percentage when analyzing amphibians (55% for forest species, 42% for open habitat species and 61% for habitat generalists). Beresford et al.8 found that AOH covered a mean of 27.6% of the range maps of 157 threatened African bird species. In 2019, Brooks et al.5 proposed a formal definition and standardized methodology to produce AOH, limiting the inputs to habitat preferences, elevation limits, and geographical range.AOH production requires knowledge of which habitat types a species occurs in and their location within the range1. Information on habitat preference is documented for each species assessed in the IUCN Red List10, following the IUCN Habitats Classification Scheme11. However, the IUCN does not define habitat classes in a spatially explicit way, therefore, we used a recently published translation table that associates IUCN Habitat Classification Scheme classes with land cover classes12. Species’ elevation limits were also extracted from the IUCN Red List.We developed AOH maps for 5,481 terrestrial mammal species and 10,651 terrestrial bird species (Fig. 1). For 1,816 bird species defined by BirdLife International as migratory, we developed separate AOH maps, for the resident, breeding, and non-breeding ranges, according to the migratory distribution of the species (Fig. 2). The maps are presented in a regular latitude/longitude grid with an approximate 100 m resolution at the equator. On average, the AOH covers 66 ± 28% of the geographical range for mammals and 64 ± 27% for birds. We used the resulting AOH maps to produce four global species richness layers for: mammals, birds, globally threatened mammals and globally threatened birds13 (Fig. 3).Fig. 1Spatial distribution maps of Tangara abbas. Maps represent (a) the geographic range21, and (b) the Area of Habitat (AOH) of the species. The AOH was produced by subtracting unsuitable habitats from the geographical range. This species’ habitats are forest and terrestrial artificial habitats and has elevation range of 0 – 1600 m.Full size imageFig. 2Spatial distribution maps of Cardellina rubrifrons, divided into resident, breeding and non-breeding areas for this migratory species. Maps represent (a) the geographic range21, and (b) the Area of Habitat (AOH) of the species. The AOH was produced by subtracting unsuitable habitats from the ranges. This species is a forest species with elevation rangelimits of 1500 – 3100 m.Full size imageFig. 3Global species richness maps for (a) terrestrial mammals (considering 5,481species) and (b) terrestrial birds (considering 10,651 species). Calculated by overlaying all species’ AOH per class, resulting inon the number of species at each grid cell, latitude/longitude grid at a resolution of 1°/1008 or approximately 100 m at the equator (EPSG:4326) with the ellipsoid WGS 1984.Full size imageThe AOH maps presented in this paper are more useful for some purposes than global species distribution models, as they reduce and standardize commissions14. They are especially useful for not well-known and wide-range species. However, we note that for well-known species alternative sources may have more accurate distributions15. Moreover, AOHs are affected by the bias and errors of the underlying data, especially relevant errors associated with documentation of species’ habitats and elevations, and the translation of habitats into land cover classes, given that habitat is a complex multidimensional concept that is challenging to match to land-cover classes12, and that the current version of the IUCN Habitat Classification Scheme on IUCN’s website is described as a draft version11.The AOH maps have multiple conservation applications5,16,17, such as assessing species’ distributions and extinction risk, improving the accuracy of conservation planning, monitoring habitat loss and fragmentation, and guiding conservation actions. AOH has been proposed as an additional spatial metric to be documented in the Red List5, and is used for the identification of Key Biodiversity Areas18. More

Study sites and study designThe study was performed in the frame of a dog ecology research project, with details on the study locations published elsewhere15,42,43. For the current study, five study sites located in Indonesia and Guatemala were included. Site selection was carried out by each country’s research team, taking into consideration rural and urban settings, as well as differing expected number of dogs present at each location. The Indonesian study sites were semi-urban Habi and rural Pogon, in the Sikka regency, at the eastern area of Flores Island (Supplementary Fig. 6). In Guatemala, the study sites were Poptún (urban setting), Sabaneta and La Romana (both rural settings), located in the Guatemalan department of Péten, in the northern part of the country (Supplementary Fig. 7). Data were collected during May to June 2018 in Guatemala and from July to September 2018 in Indonesia.In each location, a 1 km2 area was predefined using Google Earth within which the study took place. The 1 km2 area was chosen because of the research goals of another part of the project, investigating the contact network of the dogs15. Within these areas, the teams visited all dog-owning households. In each household, the study was presented to an adult of the family, who was then asked if they owned a dog and if they were willing to participate in the study. After the dog owner’s oral or written consent was granted, a questionnaire was answered, and the dogs collared. The handling of the dogs was performed by a trained veterinarian or a trained veterinary paramedic of the team.The questionnaire data was collected through interviews with the dog owners. Multiple dogs per household could be included as multiple entries in the questionnaire. The detailed questionnaire contains information on the household location, dog demographics (age, sex, reproductive status) and management (dog’s purpose, origin, confinement, vaccination status, feeding and human-mediated transportation within and outside the pre-determined area).All dogs of a household fulfilling the inclusion criteria were equipped with a geo-referenced contact sensor (GCS) developed by Bonsai Systems (https://www.bonsai-systems.com), containing a GPS module and an Ultra-High-Frequency (UHF) sensor for contact data recording43,44. GCS devices report a 5-m maximum accuracy, a run-time of up to 10 years, can store up to 4 million data points and carry a lithium-polymer-battery (LiPo). For this study, only GPS data were analysed. The GCS were set to record each dog’s geographical position at one-minute intervals. Dogs remained collared for 3 to 5 days with the duration of the data collection being limited by the device’s battery capacity, as batteries were not re-charged or changed during the study. Throughout the time of recording, date, hour, GPS coordinates and signal quality (HDOP) raw data were collected by the GPS module and amassed into the workable databases.Exclusion criteria were dogs of less than four months of age (since they were not big enough to carry a collar), sick dogs and pregnant bitches (to avoid any risk of stress-induced miscarriages). Reasons for non-participation of eligible dogs included dog owner’s absence, dog’s absence, inability to catch the dog, and refusal of participation by the dog owner. In addition, dogs foreseen for slaughtering within the following four days were excluded in Indonesia to ensure data collection for at least four to five days. All dogs included in this study were constantly free roaming or at least part-time (day only, night only and for some hours a day). Human and/or animal ethical approval were obtained depending on the country-specific regulations. All the procedures were carried out in accordance with relevant guidelines. Ethical clearance was granted in Guatemala by the UVG’s International Animal Care and Use Committee [Protocol No. I-2018(3)] and the Community Development Councils of the two rural sites, which included Maya Q’eqchi’ communities45. In Indonesia, the study was approved by the Animal Ethics Commission of the Faculty of Veterinary Medicine, Nusa Cendana University (Protocol KEH/FKH/NPEH/2019/009). In addition, dogs that participated in the study were vaccinated against rabies and/or dewormed to acknowledge the owners for their participation in the study.Data cleaningData were stored in an application developed by Bonsai Systems compatible with Apple operating system (iOS iPhone Operating Systems), downloaded as individual csv file for each unit, and further analysed in R (version 3.6.1)46.The GPS data were cleaned based on three automatised criteria. First, the speed was calculated between any two consecutive GPS fixes, and fixes with speed of  > 20 km/h were excluded, given the implausibility of a dog running at such speed over a one-minute timespan47. It is noteworthy that car travel causes speeds over 20 km/h. However, as we were interested in analysing the dog’s behaviour outside of car transports, removing these fixes was in line with our objectives. Second, the Horizontal Dilution of Precision (HDOP), which is a measure of accuracy48 and automatically recorded by the devices for each GPS fix, was used to exclude fixes with low precision. According to Lewis et al.49, GPS fixes with HDOP higher than five were excluded, which deleted 1.3% of data in Habi, 2.2% in Pogon, 3.3% in Poptún, 1.8% in La Romana and 2.1% in Sabaneta. Third, the angles built by three consecutive fixes were calculated for each dog. When studying animals’ trajectories as their measure of movement, acute inner angles are often connected to error GPS fixes50. The fixes having the 2.5% smallest angles were excluded, to target those fixes with highest risks of being errors, while balancing against the loss of GPS fixes due to the cleaning process. With the exclusion of the smallest angles, 2.6% of data were deleted in Habi, 3% in Pogon, 2.9% in Poptún, 2.6% in La Romana and 2.7% in Sabaneta. After the automatised cleaning was concluded, 18 obvious error GPS fixes (unachievable or inexplicable locations by dogs) still prevailed in the Habi dataset and were manually removed.Habitat resource identification and calculation of terrain slopeTo analyse habitat selection of the collared FRDD, resources were delimited by a 100% Minimum Convex Polygon (MCP) including all cleaned GPS fixes per study site, using QGIS51 (Fig. 1).Figure 1GPS fixes plotted over a Google satellite imagery layer with its respective outlined computed Minimum Convex Polygon (MCP) delimitating the habitat available for the study population in: (a) Habi; (b) Pogon; (c) Poptún; (d) La Romana and (e) Sabaneta. Source QGIS (version 3.4 Madeira, http://qgis.org), map data: Google Satellite.Full size imageResources were defined by taking into consideration the following criteria: resources are (i) likely to impact upon movement patterns of dogs, (ii) identifiable by landscape satellite topography, and (iii) chosen considering information on relevant gathering places for FRDD observed by the field teams. Three resources were disclosed in all study sites: buildings, roads and vegetation coverage. All habitat relevant resources were manually identified within the available area (MCP) in QGIS using satellite imagery. All building-like structures were identified using vector polygons and summed under the layer “buildings”. Roads were identified and manually traced using vector lines in all sites, except in Poptún where the roads were automatically traced using an OpenStreetMap road layer of the area (https://www.openstreetmap.org/export). A buffer vector polygon was generated to encompass the full potential width of the roads, with a 5 m width in Habi and Poptún (semi-urban and urban site) and a 2 m width in Pogon, La Romana and Sabaneta (rural sites). In Habi, a “beach” layer was defined by generating a five-meter buffer from the shoreline in both directions using a vector polygon. The layer “sea” was defined as the vector polygon resulting from the difference between the MCP sea outer limit and the beach buffer polygon. Vegetation coverage was distinct between study sites with sparse vegetation and bushes present in all sites except Pogon, and dense forest-like vegetation present in La Romana and Pogon. These two types of vegetation were defined as “low” and “high vegetation”, respectively. In Habi and La Romana, “low” and “high vegetation”, respectively, were manually identified using vector polygons and summarised under the respective layers. Finally, open field in Habi, high vegetation in Pogon and low vegetation in Poptún, La Romana and Sabaneta were the last vector layers to be established since they represented the difference between all other polygon vector layers and the MCP total area. After all resource vector polygons had been created, an encompassing vector layer was generated by merging all resource polygon vectors for final resource classification (Fig. 2). As part of the resource classification in Habi, the airport terminal and runaway as well as waterways enclosed in the MCP area were identified but excluded from the analysis.Figure 2(a) Habi, (b) Pogon, (c) Poptún, (d) La Romana and (e) Sabaneta Habitat classification vector layers. The different habitat resources, identifiable by colour, were merged to create the comprehensive Habitat classification vector. In the Indonesian sites (a, b) and Guatemalan sites (c–e) buildings are coloured red, vegetation low in Habi, Poptún, La Romana and Sabaneta is coloured light green, vegetation high in Pogon and La Romana dark green, roads black, beach yellow, sea dark blue, airport grey, waterways light blue and open field light orange. The airport area (gray) and waterways (light blue) in Habi were not classified as separate habitat layers and were excluded from further analysis. Source QGIS (version 3.4 Madeira, http://qgis.org), map data: Google Satellite.Full size imageAfter the construction of the habitat resource layers, all GPS fixes were assigned to the respective resource they were located, using the QGIS join attributes by location algorithm. Fixes located exactly on the MCP border in Indonesia were not classified automatically and had to be manually classified to the respective resource.In non-flat topographies (all locations expect Habi) we tested the hypothesis of whether the steepness would influence the dogs’ movement patterns. The degrees of slope were calculated using a 30-m raster-cell resolution (STRM 1-Arc Second Global, downloaded from the United States Geological Survey (USGS) Earth Explorer, https://earthexplorer.usgs.gov/). The slope was assigned by the QGIS join attributes by location algorithm to each GPS fix.Statistical analysisTo quantify habitat selection in each study site, we compared resources used by the dogs with the resources available, according to Freitas et al.52. Adapting the methodology applied by O’Neill et al.18, the observed number of GPS fixes for each dog was used to generate an equivalent number of locations that were randomly distributed within the MCP area using the Random points in layer bound vector tool from QGIS. For example, if dog “D300” had 100 recorded GPS fixes, 100 random points were generated within the MCP of the respective study site and assigned to “D300”. Random points were then assigned to the respective resources and slope of that location, as previously done with the observed GPS fixes. Using this approach, the habitat resources used by each dog could be compared to the available resources in the respective study site, using a regression model.Observation independence is a fundamental presupposition of any regression model. However, the spatial nature of the point-referenced data permits perception of spatial dependence. In our dataset, spatial autocorrelation was proven for all study sites using the Moran’s I test. Therefore, we applied a spatial regression model, which takes into consideration spatial autocorrelation while exploring the effects of the study variables. A mixed effects logistic regression model accounting for spatial autocorrelation was created to quantify the effect of variables on used (i.e. observed GPS fix) versus available (i.e. randomly generated GPS fixes) resources, using the fitme function in the spaMM package in R53,54. The model’s binary outcome variable was defined as either observed (1) or random (0) GPS fix, i.e. the dog being present or absent from a position. The explanatory variable was the resource classification with “buildings”, “roads”, “low vegetation”, “beach”, “sea” and “open field” as levels in Habi; “buildings”, “roads” and “high vegetation” in Pogon; “buildings”, “roads”, “low vegetation” in Poptún and Sabaneta; and “buildings”, “roads”, and “high” and “low vegetation” in La Romana. Different habitat resources were used interchangeably as reference level. In all study sites except Habi, the slope was included as an additional explanatory variable. As observations were not evenly distributed in time, with less observations recorded towards the end of the study, a variable ”hour” was added as an additional continuous fixed effect.Each observed GPS fix was assigned to the hour of its record, with the earliest timestamp registered in each study site being assigned the hour zero. The randomly generated points were randomly assigned to an hour within the determined time continuum of the observed GPS fixes. As our focus was investigating habitat selection at a population-level, we assumed there was no within-dog autocorrelation (space/time) and each dog was independent and exhibited no group behaviour38. Still, to partially account for spatial autocorrelation of each dog’s household, the random effects included in models were defined as each dog’s household geographical location recorded during fieldwork by a GPS device. The restricted maximum likelihood (REML) through Laplace approximations, which can be applied to models with non-Gaussian random effects55, and the Matérn correlation function were used to fit the spatial models with the Matérn family dispersion parameter ν, indicator of strength of decay in the spatial effect, was set at 0.554. More

Study areaFigure 2 shows the location of the study area on a map of China generated by ArcGIS software. This study’s field experiments were carried out in the Shuanghe Town agricultural comprehensive water-saving demonstration area (40°42′ N; 107°24′ E), which is located in the middle reaches of the Hetao Irrigation Area of Inner Mongolia. The duration of the experimental process ranged from April in 2018 to October in 2020. The experimental area was characterized by a mid-temperate semi-arid continental climate. The average annual precipitation was determined to be 138 mm and the average evaporation was approximately 2332 mm. The majority of the rainfall was concentrated during summer and autumn seasons, and the accumulation of salt in the surface soil was considered to be serious in the spring and winter months. The average rainfall during maize growth period was 75.3 mm. The 0 to 40 cm soil layers in the experimental area were categorized as silty loam soil, with an average bulk density ranging from 1.42 to 1.53 g cm−3. A maize straw layer with a thickness of 5 cm was buried at a depth of 40 cm, and then the land was leveled. Also, in addition to autumn watering and spring irrigation procedures, water from the Yellow River was used three times for irrigation during the entire growth period of the maize crops. The adopted irrigation method belonged to border irrigation. Urea (46% N) were used as the fertilizer types.Figure 2The location of the study area.Full size imageField trials design and data collectionWe carried out experiment 1 from 2018 to 2019, and the data obtained were used for model training and to determine the hyper-parameters. The experimental design is shown in Table 1. The PNN model trained from the data obtained in experiment 1 predicted the optimal range of irrigation amount and nitrogen application rate (N rate) for each growth period of maize. In these ranges, the soil organic matter and total nitrogen could be kept above 20 g/kg and 1.6 g/kg, respectively, the soil salt content was less than 2 g/kg, and the pH value was between 6.5 and 7.5. In order to verify the accuracy and feasibility of the range of irrigation and nitrogen application simulated by PNN, the field experiment 2 was set in 2020 based on the range simulated by PNN and to evaluate the fitting degree between measured and simulated values of soil indicators under the same amount of irrigation and nitrogen application. The experimental design is shown in Table 2.Table 1 Experimental 1 design scheme.Full size tableTable 2 Experimental 2 design scheme.Full size tableThe experimental design were repeated for three times. The plot area of each treatment measuring 8 × 9 = 72 m2. The surrounding area was separated using 1.2 m buried polyethylene plastic film, and 30 cm was left at the top to prevent fertilizer and water from flowing into each other. The field management process was consistent with that used by the local farmers. The film width of maize was 1.1 m, with each film covering two rows. The plant spacing was approximately 45 cm, and the row spacing was 35 cm. In addition, the planting density of the maize was 60,000 plants/hm2.During the entire growth period of the maize crops, soil samples were collected from the 0 to 20 cm, 20 to 40 cm, 40 to 60 cm, 60 to 80 cm, and 80 to 100 cm soil layers using a soil drill and a three-point method was adopted. The soil samples were stored at 4 °C for the determination of total nitrogen, organic matter, total salt content, and pH values. The total nitrogen, organic matter, total salt content, and pH were determined using a KDN-AA double tube azotometer, MWD-2 microwave universal digestion device, TU1810PC ultraviolet–visible spectrophotometer, and a TU18950 double beam ultraviolet–visible spectrophotometer, respectively.Soil parameters measured include organic matter (SOM), total nitrogen (TN), Salt and pH. The data set includes pre-irrigation and post-irrigation reports from 2018 to 2020. Statistical parameters regarding the soil data are shown in Table 3.Table 3 Various meteorological variables and their descriptive statistics.Full size tableThe dataset obtained in Experiment 1 in 2018 to 2019 was 2490 rows in size, the 80/20 principle was used to data into training, and testing sets were required for ML modeling; 80% of data were employed for model training, while the remaining 20% were used for testing. Specifically, the data corresponding to the treatments with the nitrogen application rate (N rate) of 75 kg/hm2 (N3) in all the treatments (W1N3, W2N3, W3N3) were used as the test set, and the data of the other treatments were used as the training set. The training set was used to initiate ML parameter training. Subsequently, The test set was employed to assess the model. The dataset size in 2020 was 1080 rows, which was used to verify ML modeling.Figure 3 shows the changes of soil indexes over time for each treatment in the field test (take the 0–40 cm soil in the main distribution area of maize roots as an example). There are differences under the influence of different irrigation amounts. When irrigation is 90 mm, soil SOM is 13.25% and 7.00% higher than 60 mm and 120 mm, and soil TN is 4.59% and 6.50% higher than 60 mm and 120 mm, respectively. The soil Salt was 23.30% lower than 60 mm, and the pH was 4.16% and 4.36% lower than that of 60 mm and 120 mm, respectively. It can be seen that irrigation of 90 mm is more favorable for increasing soil SOM and TN contents and reducing soil salinity and alkalinity. Soil SOM and TN contents were the highest at n 75 kg/hm2, which were 4.38% and 8.34% higher than those at N = 93.3 kg/hm2, respectively. Soil Salt was the lowest at N = 60 kg/hm2, which was 3.02% lower than those at N = 75 kg/hm2, with a small gap with other levels. In conclusion, nitrogen application of 75 kg/hm2 was beneficial to increase soil organic matter and nitrogen content, and nitrogen application of 60 kg/hm2 was beneficial to controlling soil salt content.Figure 3Changes in soil organic matter, total nitrogen, salinity, and pH under different treatments over time (a case study of 2019).Full size imageMachine learning (ML) models used for irrigation and nitrogen application strategiesFive ML frames were used to estimate the irrigation and N rate. These models are preference Neural Network (PNN), Support Vector Regression (SVR), Linear Regression (LR), Logistic Regression (LOR), and traditional BP Neural Networks (BPNN). Among them, the prediction effects of linear, Poly, and rbf kernel functions are respectively tried in SVR framework. The torch framework was used to train and test machine learning models in Python.Development of preference neural networkModel frameworkThe preference neural network (PNN) which was proposed for the first time in this study was a typical deep learning model. PNN can be regarded as an approximate natural function in order to describe the complete dependence of the soil fertility indexes, including the effects of soil total nitrogen, organic matter, total salt content, and pH values on irrigation and nitrogen applications. More specifically, PNN has the ability to optimize the function by constructing the mapping y = f (x, θ) and learning parameter θ.First, the input end of PNN model was defined as matrix X ∈ ℝn×d (in which n is the sample size, n = 2490; and d is the dimension of each input vector, d = 6), where {xi} i=1, …, n ∈ X represents the vectorized set of total nitrogen, organic matter, salt content, and pH used for measuring the soil fertility, as well as the nitrogen application and irrigation durations (expressed by days after sowing). At the same time, the output end of the model was defined as the matrix Y ∈ ℝn×2, which represented the levels of the irrigation and nitrogen fertilizer applications. The goal of the proposed PNN model was to learn the fixed mapping Y′ = f (X; θ) ⇒Y through the given input matrix X, where θ is the well optimized learnable parameters which can be obtained via PNN training. Meanwhile, the predicted value Y′ will infinitely approach the measured value Y. The structure and the algorithm of this study’s PNN model is shown in Fig. 4 and Table. 4.Figure 4Schematic diagram for the PNN structural connections. In the figure, it can be seen that when each input vector passed through each layer of the PNN, it is first multiplied by the Hadamard product of the weight matrix and preference value matrix for the purpose of obtaining a weight matrix with preference properties. After the matrix was activated by the Relu Function, Batch Normalization Module Methods and the Dropout Module were used for random suspension and normalization processing, and the input of the next layer was obtained.Full size imageTable 4 Algorithm of Preference neural network.Full size tableLayer-by-layer affine transformationA good definition of the affine transformation of the information flow between layers is considered to be the key to neural network model training. Generally speaking, the learnable parameter θ of each layer of a model includes the weight parameter w and the preference parameter b. The hidden representation hl of the l-th layer in PNN is defined as follows:$${h}_{l}({h}_{l-1};{W}_{l},{b}_{l})={h}_{l-1}^{mathrm{T}}{W}_{l}+{b}_{l}$$
(1)

where Wl and bl represent the learnable weight and bias variables of the l layer, respectively, and hl-1 is the hidden representation of the upper layer. Therefore, when l = 1, then h0 = X.In the present study, using the hierarchical update rules, a given input data stream was allowed to pass through each hidden layer with intermediate operations, and then finally reached the output end.Preference structureThe correlation between different production behavior factors (e.g., irrigation levels) and different natural factors (e.g., soil organic matter) differs in agricultural production. However, the traditional fully connected neural network has the characteristic that nodes of one layer are fully connected with all nodes of subsequent layers, resulting in the neurons between production behavior factors and natural factors with very weak correlation still all being connected. Conversely, connections between neurons corresponding to factors with solid correlations are not strengthened.Therefore, in this study the preference value module was specially developed. By first calculating the correlation and significance between different production behavior factors (irrigation amount, N rate) and different soil fertility factors (organic matter, total nitrogen, total salt and pH), the preference value between the above two types of variables was calculated, and the preference matrix was constructed. Then the Hadamard product of the weight matrix and preference matrix was used to realize the artificial intervention and guidance to the neural network’s learning process.In order to reduce the adverse impact of non-normality of data on correlation analysis as much as possible, this study rank-based inverse normal (RIN) transformations (i.e., conversion to rank score) methods were used to normally process the data28. The RIN transformation function used here is as follows:$$f(x)={Phi }^{-1}left(frac{{x}_{r}-frac{1}{2}}{n}right)$$
(2)

where Φ–1 is the inverse normal cumulative distribution function, and n is the sample size.The normal cumulative distribution function is represented as follows: for discrete variables, the sum of probabilities of all values less than or equal to a, and its formula is as shown below:$${F}_{X}(a)=P(Xle a)$$
(3)
The RIN normalized conversion values meet the requirements of normal distribution, Pearson correlation analysis and t-test can be directly performed, and the formula used was as follows:$$r(X,Y)=frac{mathrm{Cov}(X,Y)}{sqrt{left(mathrm{Var}left[Xright]mathrm{Var}left[mathrm{Y}right]right)}}$$
(4)

where r (X, Y) is the Pearson Correlation Coefficient, Var [X] is the variance of X, and Var [Y] is the variance of Y, Cov (X, Y) is the covariance of X and Y, which represents the overall error of the two variables. The t-test is performed on the normalized data after rank-based inverse normal (RIN) transformation method, and the formula is as follows:$$t=sqrt{frac{n-2}{1-{r}^{2}}}$$
(5)

where n is the number of samples, and r represents the Pearson Correlation Coefficient. Preference value is the concentrated embodiment of correlation and significance between variables, and the calculation formula is as follows:$${PV}_{ij}=frac{r({X}_{i},{Y}_{j})}{{P}_{ij}+e}$$
(6)

where PVij represents the preference values between the variables Xi and Yj, Xi represents the ith production behavior factor (e.g., irrigation amount), and Yj represents the jth soil fertility factor (e.g., soil organic matter content), ({P}_{ij}) is obtained by looking up the table based on the t, and e is a constant, taking 0.001 in order to prevent the denominator of the formula from being 0.In order to make the preference values of the various indicators in the same order of magnitude more stable, the preference values were normalized:$${PV}_{normal}=pm frac{left|{PV}_{i}-{PV}_{avg}right|}{sqrt{frac{sum_{i=1}^{N}{({PV}_{i}-{PV}_{avg})}^{2}}{N-1}}}$$
(7)

where N represents the number of variables related to the experimental treatments, PVi -PVavg takes the absolute value, while the positive or negative values of the PVnormal were determined by the positive or negative values of the correlation r.The PNN integrated the preference matrixes into the neural network structures by identifying the Hadamard products of the learnable weights between the preference matrixes and the input and output data. By referring to Eq. (1) in the hierarchical affine transformation, the preference constraint of PNN could be expressed as follows:$${h}_{l}({h}_{l-1};{W}_{l},{b}_{l})={h}_{l-1}^{T}{W}_{l}odot P+{b}_{l}$$
(8)

where P is the preference matrix calculated by Eq. (8), and ⊙ represents the Hadamard product of the corresponding elements of the matrix. The structure of preference neural network and preference value are shown in Figs. 5 and 6.Figure 5Schematic diagram of the preference connection structures of the preference neural networks. The depth of the network detailed in the figure only illustrates the preference connection structure (for a better demonstration), and does not indicate the depth of the PNN used in the experiment.Full size imageFigure 6PVnormal between production behavior factors and natural factors. Since soil depth, days, irrigation amount and N rate were all artificially set variables, and there was no objective correlation in the data set. Therefore, the preference values among these variables were default e = 0.001.Full size imageHyper-parameters of PNNWe conducted experiments on the datasets with varying the hyper-parameters (such as the number of PNN layers and hidden layers, the number of nodes in each layer, learning rate, dropout rate and batch size) to understand that how the Hyper-parameters impact on the performance of PNN.We select the activation function and learning rate by referring to the neural network structure commonly used in similar fields (1 hidden layer and 64 hidden nodes)29,30. It is found that ReLU has better performance than other activation functions (sigmoid, tanh). The performance is best when the learning rate is around 0.005. It is generally believed that neural networks with more hidden layers are able, with the same number of resources, to address more complex problems31, but excessively increasing network depth will easily lead to overfitting32. Since there is no direct method to select the optimal number of hidden layers and nodes33, this study first calculated the structure of one hidden layer and 64 nodes in each layer, and found that the combined effect was poor (R2 of irrigation and nitrogen application were 0.3971 and 0.4124, respectively). Therefore, the trial-and-error method is adopted. The number of hidden layers starts from 1 and is incremented by 1 to test the maximum number of 10 hidden layers. The number of nodes in each layer were tested with a maximum number of 100 hidden neurons, starting with 5 and increasing by 5.We found that when the number of hidden layers of PNN exceeds 6, and the number of nodes in each layer exceeds 65, the performance will drop significantly. The reason behind this phenomenon could be the current dataset size is insufficient for larger scale of the PNN model. In the consideration of that the size of new dataset we can obtain very year is similar to the current dataset size, we believe that current hyper-paramter settings of PNN is in a reasonable condition.After that, the number of layers was fixed as 6, and the number of nodes in each layer were tested 10 times with 60 as the starting point and 1 as the increment, we found that when the number of nodes was 64, the improvement of the fit degree was no longer noticeable. On this basis, we changed different activation functions and learning rate again, and found that PNN still has the best performance when the activation function is ReLU and the learning rate is 0.005. Then, different batch sizes and dropout rates were tried. The two parameters had weaker effects on the performance than the other parameters, and the performance was optimal at 256 and 0.1, respectively.The hyper-parameters include:

1.

number of PNN layers;

2.

number of hidden layers;

3.

types of activation function;

4.

percentage of dropout;

5.

learning rate;

6.

loss function;

7.

optimizer;

8.

batch size;

9.

number of epochs;

10.

number of workers.

The ideal PNN structure for the study comprises these layers:

1.

number of PNN layers is 8;

2.

number of hidden layers is 6;

3.

Fully connected layers with 64 nodes and ReLU activation function

4.

dropout with 0.1.

5.

the learning rate is 0.005;

6.

loss function is Huber Loss Methods (HLM);

7.

optimizer: ADAM;

8.

epochs is 500;

9.

the batch size is 256;

10.

number of workers is 6.

Hyper-parameters of other modelsLR algorithms and LOR do not have hyper-parameters that need to be adjusted. A part of the hyper-parameters of the SVR model was determined by referring to Guan Xiaoyan’s research34, and a part of the hyper-parameters of the BPNN model was determined by referring to Gu Jian’s research27. RMLP takes the same hyperparameters as PNN. The hyperparameters of SVR and BPNN models are shown in Table 5.Table 5 Hyper-parameters of other model.Full size tableModel performance evaluationThe proposed PNN model was trained and validated using the field measured data from 2020 and the performance achievements of PNN were evaluated by the root mean square errors, mean square errors, and mean absolute errors as follows:$$RMSE=sqrt{frac{{sum }_{i=1}^{n}{({y}_{ipre}-{y}_{imea})}^{2}}{n}}$$
(9)
$${R}^{2}=1-frac{{sum }_{i=1}^{n}{({y}_{ipre}-{y}_{imea})}^{2}}{{sum }_{i=1}^{n}{({y}_{ipre}-{y}_{iavg})}^{2}}$$
(10)
$$MAE=frac{{sum }_{i=1}^{n}left|{y}_{ipre}-{y}_{iavg}right|}{n}$$
(11)
Model multidimensional fertility targetsThe soil fertility grade classification of soil organic matter, soil total nitrogen content and salt content in this study was based on the soil fertility grade classification results by the Agriculture and Animal Husbandry Bureau of Bayannur City, along with the local standard Technical Specifications for the Assessment and Rating Criteria of Cultivated Land Quality (DB 15/T 1086, 2016), as the shown in Tables 6 and 7.Table 6 Soil organic matter and Soil total nitrogen degrees.Full size tableTable 7 Grading of the salinization degrees.Full size tableIn the evaluation system of soil fertility referencing the Technical Specifications for Assessment and Rating Criteria of Cultivated Land Quality (DB 15/T 1086, 2016), the pH was divided into four grades according to the membership degrees of the land productivity evaluations, as detailed in Table 8.Table 8 pH grading degrees of the cultivated land.Full size tableBased on the classification standard of soil fertility obtained by the Bureau of Agriculture and Animal Husbandry of Bayannur City, when the farmland soil is at the high fertility level, the soil organic matter and total nitrogen content should be more than 20 g/kg and 1.6 g/kg, respectively. Soil salt content was less than 2 g/kg. Meanwhile, the pH value is kept between 6.5 and 7.5. More

Vo, Q. T., Oppelt, N., Leinenkugel, P. & Kuenzer, C. Remote sensing in mapping mangrove ecosystems: An object-based approach. Remote Sens. 5, 183–201. https://doi.org/10.3390/rs5010183 (2013).Article 

Google Scholar 
Kertész, Á. & Křeček, J. Landscape degradation in the world and in Hungary. Hung. Geogr. Bull. 68, 201–221. https://doi.org/10.15201/hungeobull.68.3.1 (2019).Article 

Google Scholar 
Vorster, A. G., Evangelista, P. H., Stovall, A. E. L. & Ex, S. Variability and uncertainty in forest biomass estimates from the tree to landscape scale: The role of allometric equations. Carbon Balance Manag. 15, 8. https://doi.org/10.1186/s13021-020-00143-6 (2020).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Blackman, A. Evaluating forest conservation policies in developing countries using remote sensing data: An introduction and practical guide. For. Policy Econ. 34, 1–16. https://doi.org/10.1016/j.forpol.2013.04.006 (2013).Article 

Google Scholar 
Wilfong, B. N., Gorchov, D. L. & Henry, M. C. Detecting an invasive shrub in deciduous forest understories using remote sensing. Weed Sci. 57, 512–520. https://doi.org/10.1614/WS-09-012.1 (2009).Article 
CAS 

Google Scholar 
Dyderski, M. K. & Pawlik, Ł. Spatial distribution of tree species in mountain national parks depends on geomorphology and climate. For. Ecol. Manag. 474, 118366. https://doi.org/10.1016/j.foreco.2020.118366 (2020).Article 

Google Scholar 
Milosevic, D., Dunjić, J. & Stojanović, V. Investigating micrometeorological differences between saline steppe, forest-steppe and forest environments in northern Serbia during a clear and sunny autumn day. Geogr. Pannonica 24(3), 176–186. https://doi.org/10.5937/gp24-25885 (2020).Article 

Google Scholar 
Modzelewska, A., Fassnacht, F. E. & Stereńczak, K. Tree species identification within an extensive forest area with diverse management regimes using airborne hyperspectral data. Int. J. Appl. Earth Obs. Geoinf. 84, 101960. https://doi.org/10.1016/j.jag.2019.101960 (2020).Article 

Google Scholar 
Wulder, M. Optical remote-sensing techniques for the assessment of forest inventory and biophysical parameters. Prog. Phys. Geogr. Earth Environ. 22, 449–476. https://doi.org/10.1177/030913339802200402 (1998).Article 

Google Scholar 
Tang, L., Shao, G. & Dai, L. Roles of digital technology in China’s sustainable forestry development. Int. J. Sustain. Dev. World Ecol. 16, 94–101. https://doi.org/10.1080/13504500902794000 (2009).Article 

Google Scholar 
Richter, R., Reu, B., Wirth, C., Doktor, D. & Vohland, M. The use of airborne hyperspectral data for tree species classification in a species-rich Central European forest area. Int. J. Appl. Earth Obs. Geoinform. 52, 464–474. https://doi.org/10.1016/j.jag.2016.07.018 (2016).Article 

Google Scholar 
Thenkabail, P., Gumma, M., Teluguntla, P. & Ahmed, M. I. Hyperspectral remote sensing of vegetation and agricultural crops. Photogramm. Eng. Remote Sens. 80, 695–723 (2014).
Google Scholar 
Fassnacht, F. E. et al. Review of studies on tree species classification from remotely sensed data. Remote Sens. Environ. 186, 64–87. https://doi.org/10.1016/j.rse.2016.08.013 (2016).Article 

Google Scholar 
Vangi, E. et al. The new hyperspectral satellite PRISMA: Imagery for forest types discrimination. Sensors 21, 1182. https://doi.org/10.3390/s21041182 (2021).Article 
PubMed 
PubMed Central 

Google Scholar 
Burai, P., Beko, L., Lenart, C., Tomor, T. & Kovacs, Z. Individual tree species classification using airborne hyperspectral imagery and lidar data. In 2019 10th Workshop on Hyperspectral Imaging and Signal Processing: Evolution in Remote Sensing (WHISPERS) 1–4. https://doi.org/10.1109/WHISPERS.2019.8921016 (2019).Kumar, B., Dikshit, O., Gupta, A. & Singh, M. K. Feature extraction for hyperspectral image classification: A review. Int. J. Remote Sens. 41, 6248–6287. https://doi.org/10.1080/01431161.2020.1736732 (2020).Article 

Google Scholar 
Li, X., Li, Z., Qiu, H., Hou, G. & Fan, P. An overview of hyperspectral image feature extraction, classification methods and the methods based on small samples. Appl. Spectrosc. Rev. https://doi.org/10.1080/05704928.2021.1999252 (2021).Article 

Google Scholar 
Wang, J. & Chang, C.-I. Independent component analysis-based dimensionality reduction with applications in hyperspectral image analysis. IEEE Trans. Geosci. Remote Sens. 44, 1586–1600. https://doi.org/10.1109/TGRS.2005.863297 (2006).Article 

Google Scholar 
Hamada, Y., Stow, D. A., Coulter, L. L., Jafolla, J. C. & Hendricks, L. W. Detecting Tamarisk species (Tamarix spp.) in riparian habitats of Southern California using high spatial resolution hyperspectral imagery. Remote Sens. Environ. 109, 237–248. https://doi.org/10.1016/j.rse.2007.01.003 (2007).Article 

Google Scholar 
Ibarrola-Ulzurrun, E., Marcello, J. & Gonzalo-Martin, C. Assessment of component selection strategies in hyperspectral imagery. Entropy 19, 666. https://doi.org/10.3390/e19120666 (2017).Article 
MathSciNet 

Google Scholar 
Dabiri, Z. & Lang, S. Comparison of independent component analysis, principal component analysis, and minimum noise fraction transformation for tree species classification using APEX hyperspectral imagery. ISPRS Int. J. Geo-Inf. 7, 488. https://doi.org/10.3390/ijgi7120488 (2018).Article 

Google Scholar 
Priyadarshini, K. N., Sivashankari, V., Shekhar, S. & Balasubramani, K. Comparison and evaluation of dimensionality reduction techniques for hyperspectral data analysis. Proceedings 24, 6. https://doi.org/10.3390/IECG2019-06209 (2019).Article 

Google Scholar 
Arslan, O., Akyürek, Ö., Kaya, Ş & Şeker, D. Z. Dimension reduction methods applied to coastline extraction on hyperspectral imagery. Geocarto Int. 35, 376–390. https://doi.org/10.1080/10106049.2018.1520920 (2020).Article 

Google Scholar 
Kadavi, P. R., Lee, W.-J. & Lee, C.-W. Analysis of the pyroclastic flow deposits of mount sinabung and Merapi using landsat imagery and the artificial neural networks approach. Appl. Sci. 7, 935. https://doi.org/10.3390/app7090935 (2017).Article 

Google Scholar 
Schlosser, A. D. et al. Building extraction using orthophotos and dense point cloud derived from visual band aerial imagery based on machine learning and segmentation. Remote Sens. 12, 2397. https://doi.org/10.3390/rs12152397 (2020).Article 

Google Scholar 
Latifi, H., Fassnacht, F. & Koch, B. Forest structure modeling with combined airborne hyperspectral and LiDAR data. Remote Sens. Environ. 121, 10–25. https://doi.org/10.1016/j.rse.2012.01.015 (2012).Article 

Google Scholar 
Clark, M. L., Roberts, D. A. & Clark, D. B. Hyperspectral discrimination of tropical rain forest tree species at leaf to crown scales. Remote Sens. Environ. 96, 375–398. https://doi.org/10.1016/j.rse.2005.03.009 (2005).Article 

Google Scholar 
Melgani, F. & Bruzzone, L. Classification of hyperspectral remote sensing images with support vector machines. IEEE Trans. Geosci. Remote Sens. 42, 1778–1790. https://doi.org/10.1109/ICIECS.2009.5363456 (2004).Article 

Google Scholar 
Belgiu, M. & Drăguţ, L. Random forest in remote sensing: A review of applications and future directions. ISPRS J. Photogramm. Remote Sens. 114, 24–31. https://doi.org/10.1016/j.isprsjprs.2016.01.011 (2016).Article 

Google Scholar 
Manandhar, R., Odeh, I. O. A. & Ancev, T. Improving the accuracy of land use and land cover classification of landsat data using post-classification enhancement. Remote Sens. 1, 330–344. https://doi.org/10.3390/rs1030330 (2009).Article 

Google Scholar 
Thakkar, A. K., Desai, V. R., Patel, A. & Potdar, M. B. Post-classification corrections in improving the classification of Land Use/Land Cover of arid region using RS and GIS: The case of Arjuni watershed, Gujarat, India. Egypt. J. Remote Sens. Space Sci. 20, 79–89. https://doi.org/10.1016/j.ejrs.2016.11.006 (2017).Article 

Google Scholar 
El-Hattab, M. M. Applying post classification change detection technique to monitor an Egyptian coastal zone (Abu Qir Bay), Egypt. J. Remote Sens. Space Sci. 19, 23–36. https://doi.org/10.1016/j.ejrs.2016.02.002 (2016).Article 

Google Scholar 
Bhosale, N., Manza, R., Kale, K., Scholar, R. & Professor, A. Analysis of effect of gaussian, salt and pepper noise removal from noisy remote sensing images. Pceedings of teh Second International Conference on ERCICA 386–390. http://rameshmanza.in/Publication/Narayan_Bhosle/Analysis%20of%20Effect%20of%20Gaussian.pdf (2014).Schöll, K., Kiss, A., Dinka, M. & Berczik, Á. Flood-pulse effects on zooplankton assemblages in a river-floodplain system (Gemenc Floodplain of the Danube, Hungary). Int. Rev. Hydrobiol. 97, 41–54. https://doi.org/10.1002/iroh.201111427 (2012).Article 

Google Scholar 
Ágoston-Szabó, E., Schöll, K., Kiss, A. & Dinka, M. The effects of tree species richness and composition on leaf litter decomposition in a Danube oxbow lake (Gemenc, Hungary). Fundam. Appl. Limnol. https://doi.org/10.1127/fal/2017/0675 (2017).Article 

Google Scholar 
Guti, G. Water bodies in the Gemenc floodplain of the Danube, Hungary: (A theoretical basis for their typology). Opusc Zool. 33, 49–60 (2001).
Google Scholar 
Berczik, Á. & Dinka, M. Bibliography of hydrobiological research on the Gemenc and Béda: Karapancsa floodplains of the River Danube (1498–1436 rkm) including the publications of the Danube Research Institute of the Hungarian Academy of Sciences between 1968 and 2017. Opusc. Zool. 49, 191–197. https://doi.org/10.18348/opzool.2018.2.191 (2018).Article 

Google Scholar 
Ceulemans, R., McDonald, A. J. S. & Pereira, J. S. A comparison among eucalypt, poplar and willow characteristics with particular reference to a coppice, growth-modelling approach. Biomass Bioenergy 11, 215–231. https://doi.org/10.1016/0961-9534(96)00035-9 (1996).Article 

Google Scholar 
Haneca, K., Katarina, Č & Beeckman, H. Oaks, tree-rings and wooden cultural heritage: A review of the main characteristics and applications of oak dendrochronology in Europe. J. Archaeol. Sci. 36, 1–11. https://doi.org/10.1016/j.jas.2008.07.005 (2009).Article 

Google Scholar 
Jones, T. G., Coops, N. C. & Sharma, T. Assessing the utility of airborne hyperspectral and LiDAR data for species distribution mapping in the coastal Pacific Northwest, Canada. Remote Sens. Environ. 114, 2841–2852. https://doi.org/10.1016/j.rse.2010.07.002 (2010).Article 

Google Scholar 
Sothe, C. et al. Tree species classification in a highly diverse subtropical forest integrating UAV-based photogrammetric point cloud and hyperspectral data. Remote Sens. 11, 1338. https://doi.org/10.3390/rs11111338 (2019).Article 

Google Scholar 
Nambiar, E. K. S. & Sands, R. Competition for water and nutrients in forests. Can. J. For. Res. 23, 1955–1968. https://doi.org/10.1139/x93-247 (1993).Article 

Google Scholar 
Mayoral, C., Calama, R., Sánchez-González, M. & Pardos, M. Modelling the influence of light, water and temperature on photosynthesis in young trees of mixed Mediterranean forests. New For. 46, 485–506. https://doi.org/10.1007/s11056-015-9471-y (2015).Article 

Google Scholar 
Stojanović, D. B., Levanič, T., Matović, B. & Orlović, S. Growth decrease and mortality of oak floodplain forests as a response to change of water regime and climate. Eur. J. For. Res. 134, 555–567. https://doi.org/10.1007/s10342-015-0871-5 (2015).Article 

Google Scholar 
Dyderski, M. K. & Jagodziński, A. M. Impact of invasive tree species on natural regeneration species composition, diversity, and density. Forests 11, 456. https://doi.org/10.3390/f11040456 (2020).Article 

Google Scholar 
Jia, S., Ji, Z., Qian, Y. & Shen, L. Unsupervised band selection for hyperspectral imagery classification without manual band removal. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens. 5, 531–543. https://doi.org/10.1109/JSTARS.2012.2187434 (2012).Article 

Google Scholar 
Karpouzli, E. & Malthus, T. The empirical line method for the atmospheric correction of IKONOS imagery. Int. J. Remote Sens. 24, 1143–1150. https://doi.org/10.1080/0143116021000026779 (2003).Article 

Google Scholar 
Richards, J. A. Remote Sensing Digital Image Analysis (Springer, 2013). https://doi.org/10.1007/978-3-642-30062-2.Book 

Google Scholar 
Sharifi Hashjin, S. & Khazai, S. A new method to detect targets in hyperspectral images based on principal component analysis. Geocarto Int. 37, 2679–2697. https://doi.org/10.1080/10106049.2020.1831625 (2022).Article 

Google Scholar 
Kaiser, H. F. The varimax criterion for analytic rotation in factor analysis. Psychometrika 23, 187–200 (1958).Article 
MATH 

Google Scholar 
Shah, C. A., Arora, M. K. & Varshney, P. K. Unsupervised classification of hyperspectral data: An ICA mixture model based approach. Int. J. Remote Sens. 25, 481–487. https://doi.org/10.1080/01431160310001618040 (2004).Article 

Google Scholar 
Tharwat, A. Independent component analysis: An introduction. Appl. Comput. Inform. 17, 222–249. https://doi.org/10.1016/S1364-6613(00)01813-1 (2020).Article 

Google Scholar 
Villa, A., Chanussot, J., Jutten, C., Benediktsson, J. A. & Moussaoui, S. On the use of ICA for hyperspectral image analysis. In 2009 IEEE International Geoscience and Remote Sensing Symposium vol. 4 IV-97-IV–100. https://doi.org/10.1109/IGARSS.2009.5417363 (2009).Hyvärinen, A. & Oja, E. Independent component analysis: Algorithms and applications. Neural Netw. 13, 411–430. https://doi.org/10.1016/s0893-6080(00)00026-5 (2000).Article 
PubMed 

Google Scholar 
Otukei, J. R. & Blaschke, T. Land cover change assessment using decision trees, support vector machines and maximum likelihood classification algorithms. Int. J. Appl. Earth Obs. Geoinf. 12, S27–S31. https://doi.org/10.1016/j.jag.2009.11.002 (2010).Article 

Google Scholar 
Murty, M. N. & Raghava, R. Kernel-based SVM. In Support Vector Machines and Perceptrons: Learning, Optimization, Classification, and Application to Social Networks (eds Murty, M. N. & Raghava, R.) 57–67 (Springer, 2016). https://doi.org/10.1007/978-3-319-41063-0_5.Chapter 
MATH 

Google Scholar 
Seidl, D., Ružiak, I., Koštialová Jančíková, Z. & Koštial, P. Sensitivity analysis: A tool for tailoring environmentally friendly materials. Expert Syst. Appl. 208, 118039. https://doi.org/10.1016/j.eswa.2022.118039 (2022).Article 

Google Scholar 
Zhao, D., Pang, Y., Liu, L. & Li, Z. Individual tree classification using airborne LiDAR and hyperspectral data in a natural mixed forest of Northeast China. Forests 11, 303. https://doi.org/10.3390/f11030303 (2020).Article 

Google Scholar 
Aksoy, S. & Akcay, H. G. Multi-resolution segmentation and shape analysis for remote sensing image classification. In Proceedings of 2nd International Conference on Recent Advances in Space Technologies, 2005. RAST 2005. 599–604 (2005). https://doi.org/10.1109/RAST.2005.1512638.Dalponte, M., Ørka, H. O., Ene, L. T., Gobakken, T. & Næsset, E. Tree crown delineation and tree species classification in boreal forests using hyperspectral and ALS data. Remote Sens. Environ. 140, 306–317. https://doi.org/10.1016/j.rse.2013.09.006 (2014).Article 

Google Scholar 
Amini, S., Homayouni, S., Safari, A. & Darvishsefat, A. A. Object-based classification of hyperspectral data using Random Forest algorithm. Geo-Spat. Inf. Sci. 21, 127–138. https://doi.org/10.1080/10095020.2017.1399674 (2018).Article 

Google Scholar 
Congalton, R. G. A review of assessing the accuracy of classifications of remotely sensed data. Remote Sens. Environ. 37, 35–46. https://doi.org/10.1016/0034-4257(91)90048-B (1991).Article 

Google Scholar 
Foody, G. M. Status of land cover classification accuracy assessment. Remote Sens. Environ. 80, 185–201. https://doi.org/10.1016/S0034-4257(01)00295-4 (2002).Article 

Google Scholar 
Tharwat, A. Classification assessment methods. Appl. Comput. Inform. 17, 168–192. https://doi.org/10.1016/j.aci.2018.08.003 (2020).Article 

Google Scholar 
Field, F. Discovering Statistics Using IBM SPSS Statistics. SAGE Publications Ltd https://uk.sagepub.com/en-gb/eur/discovering-statistics-using-ibm-spss-statistics/book257672 (2022).R Core Team. R: A language and environment for statistical computing. https://www.gbif.org/tool/81287/r-a-language-and-environment-for-statistical-computing (2022).Galucci, M. Generalized Mixed Models module. R package version 2.0.5. https://gamlj.github.io/gzlmmixed.html More

United Nations. World Urbanization Prospects: The 2018 revision. (Department of Economic and Social Affairs, Population Division, United Nations, 2018).Grimm, N. B. et al. Global change and the ecology of cities. Science 319, 756–760 (2008).Article 

Google Scholar 
McPhearson, T. et al. Advancing urban ecology toward a science of cities. Bioscience 66, 198–212 (2016).Article 

Google Scholar 
Dodman, D. et al. Cities, settlements and key infrastructure. In Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (eds. Pörtner, H.-O. et al.) 997–1040 (Cambridge University Press, Cambridge, UK and New York, NY, USA, 2022).Díaz, S. et al. Assessing nature’s contributions to people: Recognizing culture, and diverse sources of knowledge, can improve assessments. Science 359, 270–272 (2018).Article 

Google Scholar 
Grabowski, Z. J., McPhearson, T., Matsler, A. M., Groffman, P. & Pickett, S. T. A. What is green infrastructure? A study of definitions in US city planning. Front. Ecol. Environ. 20, 152–160 (2022).Article 

Google Scholar 
Childers, D. L. et al. Urban ecological infrastructure: An inclusive concept for the non-built urban environment. Elementa 7, 1–14 (2019).
Google Scholar 
Gómez-Baggethun, E. et al. Urban ecosystem services. In Urbanization, Biodiversity and Ecosystem Services: Challenges and Opportunities (eds. Elmqvist, T. et al.) 175–251 (Springer, Netherlands, 2013).Díaz, S. & Cabido, M. Vive la différence: Plant functional diversity matters to ecosystem processes. Trends Ecol. Evol. 16, 646–655 (2001).Article 

Google Scholar 
Burkhard, B. & Maes, J. Mapping Ecosystem Services (Pensoft Publishers, Sofia, 2017).Eviner, V. T. & Chapin, F. S. Functional Matrix: A conceptual framework for predicting multiple plant effects on ecosystem processes. Annu. Rev. Ecol. Evol. Syst. 34, 455–485 (2003).Article 

Google Scholar 
Lavorel, S., McIntyre, S., Landsberg, J. & Forbes, T. D. A. Plant functional classifications: From general groups to specific groups based on response to disturbance. Trends Ecol. Evol. 12, 474–478 (1997).Article 

Google Scholar 
Cornelissen, J. H. C. et al. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust. J. Bot. 51, 335–380 (2003).Article 

Google Scholar 
Suding, K. N. et al. Scaling environmental change through the community-level: A trait-based response-and-effect framework for plants. Glob. Chang. Biol. 14, 1125–1140 (2008).Article 

Google Scholar 
Lavorel, S. & Garnier, E. Predicting changes in community composition and ecosystem functioning from plant traits: Revisiting the Holy Grail. Funct. Ecol. 16, 545–556 (2002).Article 

Google Scholar 
Hevia, V. et al. Trait-based approaches to analyze links between the drivers of change and ecosystem services: Synthesizing existing evidence and future challenges. Ecol. Evol. 7, 831–844 (2017).Article 

Google Scholar 
Cadotte, M. W., Carscadden, K. & Mirotchnick, N. Beyond species: Functional diversity and the maintenance of ecological processes and services. J. Appl. Ecol. 48, 1079–1087 (2011).Article 

Google Scholar 
Lavorel, S. Plant functional effects on ecosystem services. J. Ecol. 101, 4–8 (2013).Article 

Google Scholar 
Andersson, E. et al. What are the traits of a social-ecological system: Towards a framework in support of urban sustainability. npj Urban Sustain. 1, 14 (2021).Article 

Google Scholar 
Pickett, S. T. A. et al. Urban ecological systems: Linking terrestrial ecological, physical, and socioeconomic components of metropolitan areas. Annu. Rev. Ecol. Syst. 32, 127–157 (2001).Article 

Google Scholar 
McPhearson, T., Haase, D., Kabisch, N. & Gren, Å. Advancing understanding of the complex nature of urban systems. Ecol. Indic. 70, 566–573 (2016).Article 

Google Scholar 
Zhou, W., Pickett, S. T. A. & McPhearson, T. Conceptual frameworks facilitate integration for transdisciplinary urban science. npj Urban Sustain. 1, 1 (2021).Article 

Google Scholar 
Andersson, E. et al. Scale and context dependence of ecosystem service providing units. Ecosyst. Serv. 12, 157–164 (2015).Article 

Google Scholar 
Pinho, P. et al. Research agenda on biodiversity and ecosystem functions and services in European cities. Basic Appl. Ecol. 53, 124–133 (2021).Article 

Google Scholar 
Bullock, J. M. et al. Human-mediated dispersal and the rewiring of spatial networks. Trends Ecol. Evol. 33, 958–970 (2018).Article 

Google Scholar 
Avolio, M. L., Swan, C., Pataki, D. E. & Jenerette, G. D. Incorporating human behaviors into theories of urban community assembly and species coexistence. Oikos 130, 1849–1864 (2021).Article 

Google Scholar 
Aronson, M. F. J. et al. Hierarchical filters determine community assembly of urban species pools. Ecology 97, 2952–2963 (2016).Article 

Google Scholar 
Woodward, F. I. & Diament, A. D. Functional approaches to predicting the ecological effects of global change. Funct. Ecol. 5, 212 (1991).Article 

Google Scholar 
Diaz, S., Cabido, M. & Casanoves, F. Plant functional traits and environmental filters at a regional scale. J. Veg. Sci. 9, 113–122 (1998).Article 

Google Scholar 
Boet, O., Arnan, X. & Retana, J. The role of environmental vs. biotic filtering in the structure of European ant communities: A matter of trait type and spatial scale. PLoS ONE 15, e0228625 (2020).Article 

Google Scholar 
Grimm, N. B., Grove, J. M., Pickett, S. T. A. & Redman, C. L. Integrated approaches to long-term studies of urban ecological systems. Bioscience 50, 571–584 (2000).Article 

Google Scholar 
Vandewalle, M. et al. Functional traits as indicators of biodiversity response to land use changes across ecosystems and organisms. Biodivers. Conserv. 19, 2921–2947 (2010).Article 

Google Scholar 
Williams, N. S. G. et al. A conceptual framework for predicting the effects of urban environments on floras. J. Ecol. 97, 4–9 (2009).Article 

Google Scholar 
Cavender-Bares, J. et al. Horticultural availability and homeowner preferences drive plant diversity and composition in urban yards. Ecol. Appl. 30, 1–16 (2020).Article 

Google Scholar 
Pearse, W. D. et al. Homogenization of plant diversity, composition, and structure in North American urban yards. Ecosphere 9, e02105 (2018).Article 

Google Scholar 
Cubino, J. P. et al. Drivers of plant species richness and phylogenetic composition in urban yards at the continental scale. Landsc. Ecol. 34, 63–77 (2019).Article 

Google Scholar 
Oke, T. R. The energetic basis of the urban heat island. Q. J. R. Meteorol. Soc. 108, 1–24 (1982).
Google Scholar 
Sukopp, H. Human-caused impact on preserved vegetation. Landsc. Urban Plan. 68, 347–355 (2004).Article 

Google Scholar 
Díaz, S. et al. Incorporating plant functional diversity effects in ecosystem service assessments. Proc. Natl. Acad. Sci. USA 104, 20684–20689 (2007).Article 

Google Scholar 
Williams, N. S. G., Hahs, A. K. & Vesk, P. A. Urbanisation, plant traits and the composition of urban floras. Perspect. Plant Ecol. Evol. Syst. 17, 78–86 (2015).Article 

Google Scholar 
Teskey, R. et al. Responses of tree species to heat waves and extreme heat events. Plant Cell Environ. 38, 1699–1712 (2015).Article 

Google Scholar 
Jochner, S. & Menzel, A. Urban phenological studies—past, present, future. Environ. Pollut. 203, 250–261 (2015).Article 

Google Scholar 
Cleland, E. E., Chuine, I., Menzel, A., Mooney, H. A. & Schwartz, M. D. Shifting plant phenology in response to global change. Trends Ecol. Evol. 22, 357–365 (2007).Article 

Google Scholar 
de Bello, F. et al. Towards an assessment of multiple ecosystem processes and services via functional traits. Biodivers. Conserv. 19, 2873–2893 (2010).Article 

Google Scholar 
Santangelo, J. S. et al. Global urban environmental change drives adaptation in white clover. Science 375, 1275–1281 (2022).Article 

Google Scholar 
Blois, J. L., Zarnetske, P. L., Fitzpatrick, M. C. & Finnegan, S. Climate change and the past, present, and future of biotic interactions. Science 341, 499–504 (2013).Article 

Google Scholar 
Martin, C. A., Warren, P. S. & Kinzig, A. P. Neighborhood socioeconomic status is a useful predictor of perennial landscape vegetation in residential neighborhoods and embedded small parks of Phoenix, AZ. Landsc. Urban Plan. 69, 355–368 (2004).Article 

Google Scholar 
Kinzig, A. P., Warren, P., Martin, C., Hope, D. & Katti, M. The effects of human socioeconomic status and cultural characteristics on urban patterns of biodiversity. Ecol. Soc. 10, 23 (2005).Article 

Google Scholar 
Stephenson, J. The cultural values model: An integrated approach to values in landscapes. Landsc. Urban Plan. 84, 127–139 (2008).Article 

Google Scholar 
Andersson, E., Barthel, S. & Ahrné, K. Measuring social–ecological dynamics behind the generation of ecosystem services. Ecol. Appl. 17, 1267–1278 (2007).Article 

Google Scholar 
Fraser, E. D. G. & Kenney, W. A. Cultural background and landscape history as factors affecting perceptions of the urban forest. J. Arboric. 26, 106–113 (2000).
Google Scholar 
Hope, D. et al. Socioeconomics drive urban plant diversity. Proc. Natl. Acad. Sci. USA 100, 8788–8792 (2003).Article 

Google Scholar 
Avolio, M. L. et al. Understanding preferences for tree attributes: The relative effects of socio-economic and local environmental factors. Urban Ecosyst. 18, 73–86 (2015).Article 

Google Scholar 
Körmöndi, B., Tempfli, J., Kocsis, J. B., Adams, J. & Szkordilisz, F. E. The secret ingredient—The role of governance in green infrastructure development: Through the examples of European cities. IOP Conf. Ser. Earth Environ. Sci. 323, (2019).Conway, T. M. & Vander Vecht, J. Growing a diverse urban forest: species selection decisions by practitioners planting and supplying trees. Landsc. Urban Plan. 138, 1–10 (2015).Article 

Google Scholar 
Lack, W. H. The Book of Palms (Taschen-Bibliotheca Universalis, 2015).Grilo, F. et al. Using green to cool the grey: Modelling the cooling effect of green spaces with a high spatial resolution. Sci. Total Environ. 724, 138182 (2020).Article 

Google Scholar 
Prasifka, J. R. et al. Using nectar-related traits to enhance crop–pollinator interactions. Front. Plant Sci. 9, 1–8 (2018).Article 

Google Scholar 
Veerkamp, C. J. et al. A review of studies assessing ecosystem services provided by urban green and blue infrastructure. Ecosyst. Serv. 52, 101367 (2021).Article 

Google Scholar 
Theodorou, P. et al. Urban areas as hotspots for bees and pollination but not a panacea for all insects. Nat. Commun. 11, 576 (2020).Article 

Google Scholar 
Farmer, J. Trees in Paradise: A California History (WW Norton & Company, 2013).Goodness, J., Andersson, E., Anderson, P. M. L. & Elmqvist, T. Exploring the links between functional traits and cultural ecosystem services to enhance urban ecosystem management. Ecol. Indic. 70, 597–605 (2016).Article 

Google Scholar 
Masterson, V. A. et al. The contribution of sense of place to social-ecological systems research: A review and research agenda. Ecol. Soc. 22, 49 (2017).Article 

Google Scholar 
Masterson, V. A., Enqvist, J. P., Stedman, R. C. & Tengö, M. Sense of place in social-ecological systems: From theory to empirics. Sustain. Sci. 14, 555–564 (2019).Article 

Google Scholar 
Mukherjee, A. & Agrawal, M. Use of GLM approach to assess the responses of tropical trees to urban air pollution in relation to leaf functional traits and tree characteristics. Ecotoxicol. Environ. Saf. 152, 42–54 (2018).Article 

Google Scholar 
Singh, S. K., Rao, D. N., Agrawal, M., Pandey, J. & Naryan, D. Air pollution tolerance index of plants. J. Environ. Manage. 32, 45–55 (1991).Article 

Google Scholar 
Mukherjee, A. & Agrawal, M. Pollution response score of tree species in relation to ambient air quality in an urban area. Bull. Environ. Contam. Toxicol. 96, 197–202 (2016).Article 

Google Scholar 
Barwise, Y. & Kumar, P. Designing vegetation barriers for urban air pollution abatement: A practical review for appropriate plant species selection. npj Clim. Atmos. Sci. 3, 12 (2020).Article 

Google Scholar 
Grote, R. et al. Functional traits of urban trees: Air pollution mitigation potential. Front. Ecol. Environ. 14, 543–550 (2016).Article 

Google Scholar 
Tomson, M. et al. Green infrastructure for air quality improvement in street canyons. Environ. Int. 146, 106288 (2021).Article 

Google Scholar  More

Rogers, A. et al. Antarctic futures: An assessment of climate-driven changes in ecosystem structure, function, and service provisioning in the southern ocean. Annual Review of Marine Science 12, 87–120, https://doi.org/10.1146/annurev-marine-010419-011028 (2020).Article 
PubMed 

Google Scholar 
Tin, T. et al. Impacts of local human activities on the antarctic environment. Antarctic Science 21, 3–33, https://doi.org/10.1017/S0954102009001722 (2009).Article 

Google Scholar 
Pineda-Metz, S. E. A., Gerdes, D. & Richter, C. Benthic fauna declined on a whitening antarctic continental shelf. Nature Communications 11, 2226, https://doi.org/10.1038/s41467-020-16093-z (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Convey, P. Antarctic terrestrial biodiversity in a changing world. Polar Biology 34, 1629, https://doi.org/10.1007/s00300-011-1068-0 (2011).Article 

Google Scholar 
Kang, Y. H. et al. Composition and structure of the marine benthic community in terra nova bay, antarctica: Responses of the benthic assemblage to disturbances. PLOS ONE 14, 1–16, https://doi.org/10.1371/journal.pone.0225551 (2019).Article 

Google Scholar 
Piazza, P. et al. Underwater photogrammetry in antarctica: long-term observations in benthic ecosystems and legacy data rescue. Polar Biology 42, 1061–1079, https://doi.org/10.1007/s00300-019-02480-w (2019).Article 

Google Scholar 
SOOS. Southern Ocean Observing System – Report on the 2017 Ross Sea Working Group Meeting. http://www.soos.aq. [Online; accessed 2022/11/15] (2017).SCAR. Scientific Committee on Antarctic Research. https://www.scar.org. [Online; accessed 2022/11/15] (2021).ANTOS. Antarctic near-shore and terrestrial observing system. https://www.scar.org/science/antos/home. [Online; accessed 2022/11/15] (2015).Dayton, P. K. et al. Benthic responses to an antarctic regime shift: food particle size and recruitment biology. Ecological Applications 29, e01823, https://doi.org/10.1002/eap.1823 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Watters, G. M., Hinke, J. T. & Reiss, C. S. Long-term observations from antarctica demonstrate that mismatched scales of fisheries management and predator-prey interaction lead to erroneous conclusions about precaution. Scientific Reports 10, 2314, https://doi.org/10.1038/s41598-020-59223-9 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Bolinesi, F. et al. Spatial-related community structure and dynamics in phytoplankton of the ross sea, antarctica. Frontiers in Marine Science 7, https://doi.org/10.3389/fmars.2020.574963 (2020).Stenni, B. et al. Three-year monitoring of stable isotopes of precipitation at concordia station, east antarctica. The Cryosphere 10, 2415–2428, https://doi.org/10.5194/tc-10-2415-2016 (2016).Article 

Google Scholar 
Ramesh, K. & Soni, V. Perspectives of antarctic weather monitoring and research efforts. Polar Science 18, 183–188, https://doi.org/10.1016/j.polar.2018.04.005 (2018). Recent Advances in Climate Science of Polar Region (to commemorate the contributions of Late Dr. S.Z. Qasim, a pioneering doyen of the Indian Polar programme).Article 

Google Scholar 
Shepherd, A. et al. Mass balance of the antarctic ice sheet from 1992 to 2017. Nature 558, 219–222, https://doi.org/10.1038/s41586-018-0179-y (2018).Article 

Google Scholar 
Budge, J. S. & Long, D. G. A comprehensive database for antarctic iceberg tracking using scatterometer data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 11, 434–442, https://doi.org/10.1109/JSTARS.2017.2784186 (2018).Article 

Google Scholar 
Rignot, E. et al. Four decades of antarctic ice sheet mass balance from 1979–2017. Proceedings of the National Academy of Sciences of the United States of America 116, 1095–1103, https://doi.org/10.1073/pnas.1812883116 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Barbat, M. M., Rackow, T., Wesche, C., Hellmer, H. H. & Mata, M. M. Automated iceberg tracking with a machine learning approach applied to sar imagery: A weddell sea case study. ISPRS Journal of Photogrammetry and Remote Sensing 172, 189–206, https://doi.org/10.1016/j.isprsjprs.2020.12.006 (2021).Article 

Google Scholar 
Aguzzi, J. et al. New high-tech flexible networks for the monitoring of deep-sea ecosystems. Environmental Science & Technology 53, 6616–6631, https://doi.org/10.1021/acs.est.9b00409 (2019).Article 

Google Scholar 
Piazza, P., Gattone, S., Guzzi, A. & Schiaparelli, S. Towards a robust baseline for long-term monitoring of antarctic coastal benthos. Hydrobiologia 847, 1753–1771, https://doi.org/10.1007/s10750-020-04177-2 (2020).Article 

Google Scholar 
Rountree, R. et al. Towards an optimal design for ecosystem-level ocean observatories. Oceanography and Marine Biology 58, 79–105, https://doi.org/10.1201/9780429351495-2 (2020).Article 

Google Scholar 
Katsanevakis, S. et al. Monitoring marine populations and communities: Methods dealing with imperfect detectability. Aquatic Biology 16, 31–52, https://doi.org/10.3354/ab00426 (2012).Article 

Google Scholar 
Zampoukas, N. et al. Technical guidance on monitoring for the marine strategy framework directive. Tech. Rep., European Commission, Report EUR 26499 (2014).Bicknell, A. W., Godley, B. J., Sheehan, E. V., Votier, S. C. & Witt, M. J. Camera technology for monitoring marine biodiversity and human impact. Frontiers in Ecology and the Environment 14, 424–432, https://doi.org/10.1002/fee.1322 (2016).Article 

Google Scholar 
European Marine Board. Working Group on Big Data in Marine Science. https://www.marineboard.eu/publications/big-data-marine-science. [Online; accessed 2022/11/15] (2020).Zurowietz, M. & Nattkemper, T. W. Current trends and future directions of large scale image and video annotation: Observations from four years of biigle 2.0. Frontiers in Marine Science 8, https://doi.org/10.3389/fmars.2021.760036 (2021).Kim, S. L., Thurber, A., Hammerstrom, K. & Conlan, K. Seastar response to organic enrichment in an oligotrophic polar habitat. Journal of Experimental Marine Biology and Ecology 346, 66–75, https://doi.org/10.1016/j.jembe.2007.03.004 (2007).Article 

Google Scholar 
Peirano, A., Bordone, A., Marini, S., Piazza, P. & Schiaparelli, S. A simple time-lapse apparatus for monitoring macrozoobenthos activity in antarctica. Antarctic Science 28, 473–474, https://doi.org/10.1017/S0954102016000377 (2016).Article 

Google Scholar 
Peirano, A., Marini, S., Bordone, A. & Schiaparelli, S. ICE-LAPSE: Analysis of antarctic benthos dynamics by using non-destructive monitoring devices and permanent stations, pnra 2013/az1.16, funded by the italian national antarctic program (2015-2016).Marini, S. et al. Long-term automated visual monitoring of antarctic benthic fauna. Methods in Ecology and Evolution 13, 1746–1764, https://doi.org/10.1111/2041-210X.13898 (2022).Article 

Google Scholar 
Marini, S. et al. EP2863257 (A1) – Underwater images acquisition and processing system. https://data.epo.org/gpi/EP2863257B1. [Online; accessed 2022/11/15] (2013).Corgnati, L. et al. Looking inside the ocean: Toward an autonomous imaging system for monitoring gelatinous zooplankton. Sensors 16, 2124, https://doi.org/10.3390/s16122124 (2016).Article 
PubMed 
PubMed Central 

Google Scholar 
Marini, S. et al. Automated estimate of fish abundance through the autonomous imaging device guard1. Measurement 126, 72–75, https://doi.org/10.1016/j.measurement.2018.05.035 (2018).Article 

Google Scholar 
Pensieri, S. et al. Environmental acoustic noise observations in tethys bay (terra nova bay, ross sea, antarctica). In 2014 Oceans – St. John’s, 1–6, https://doi.org/10.1109/OCEANS.2014.7003196 (2014).Jung, J. et al. Multibeam bathymetry and distribution of clay minerals on surface sediments of a small bay in terra nova bay, antarctica. Minerals 11, https://doi.org/10.3390/min11010072 (2021).Balog, I. et al. Estimation of direct normal irradiance at antarctica for concentrated solar technology. Applied System Innovation 2, https://doi.org/10.3390/asi2030021 (2019).Caputi, S. S. et al. Seasonal food web dynamics in the antarctic benthos of tethys bay (ross sea): Implications for biodiversity persistence under different seasonal sea-ice coverage. Frontiers in Marine Science 7, 1046, https://doi.org/10.3389/fmars.2020.594454 (2020).Article 

Google Scholar 
van Leeuwe, M. A. et al. Annual patterns in phytoplankton phenology in antarctic coastal waters explained by environmental drivers. Limnology and Oceanography 65, 1651–1668, https://doi.org/10.1002/lno.11477 (2020).Article 

Google Scholar 
OEngineering. OEngineering s.r.l. – GUARD-1, Underwater Autonomous Smart Camera. https://www.oengineering.eu//GUARD-1/. [Online; accessed 2022/11/15] (2021).Magic Lantern. https://magiclantern.fm. [Online; accessed 2022/11/15] (2021).Marini, S. et al. Guard1: An autonomous system for gelatinous zooplankton image-based recognition. In OCEANS 2015 – Genova, 1–7, https://doi.org/10.1109/OCEANS-Genova.2015.7271704 (2015).CR2. The Canon RAW (CRW) File Format. https://exiftool.org/canon_raw.html. [Online; accessed 2022/11/15] (2022).Marini, S. et al. ICE-LAPSE image dataset. Zenodo https://doi.org/10.5281/zenodo.6418163 (2022).LabelImg. A graphical image annotation tool. https://github.com/tzutalin/labelImg. [Online; accessed 2022/11/15] (2021).Schoening, T. et al. Making marine image data fair. Scientific Data 9, 414, https://doi.org/10.1038/s41597-022-01491-3 (2022).Article 
PubMed 
PubMed Central 

Google Scholar 
Cattaneo-Vietti, R., Chiantore, M., Schiaparelli, S. & Albertelli, G. Shallow- and deep-water mollusc distribution at terra nova bay (ross sea, antarctica). Polar Biology 23, 173–182, https://doi.org/10.1007/s003000050024 (2000).Article 

Google Scholar 
Cattaneo-Vietti, R. et al. Spatial and Vertical Distribution of Benthic Littoral Communities in Terra Nova Bay, 503–514 (Springer Berlin Heidelberg, Berlin, Heidelberg, 2000).Cummings, V. J. et al. Linking ross sea coastal benthic communities to environmental conditions: Documenting baselines in a spatially variable and changing world. Frontiers in Marine Science 5, 232, https://doi.org/10.3389/fmars.2018.00232 (2018).Article 

Google Scholar 
Redmon, J., Divvala, S., Girshick, R. & Farhadi, A. You only look once: Unified, real-time object detection. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 779–788, https://doi.org/10.1109/CVPR.2016.91 (2016).YOLO V5. https://github.com/ultralytics/yolov5. [Online; accessed 2022/11/15] (2022). More

Field, C. B., Behrenfeld, M. J., Randerson, J. T. & Falkowski, P. Primary production of the biosphere: Integrating terrestrial and oceanic components. Science 281, 237–240. https://doi.org/10.1126/science.281.5374.237 (1998).Article 
CAS 
PubMed 

Google Scholar 
Valladares, F. In Progress in Botany Vol. 64 (eds Esser, K. et al.) 439–471 (Springer, 2003).Chapter 

Google Scholar 
Anthony, K. R. N., Ridd, P. V., Orpin, A. R., Larcombe, P. & Lough, J. Temporal variation of light availability in coastal benthic habitats: Effects of clouds, turbidity, and tides. Limnol. Oceanogr. 49, 2201–2211. https://doi.org/10.4319/lo.2004.49.6.2201 (2004).Article 

Google Scholar 
Gattuso, J. P. et al. Light availability in the coastal ocean: Impact on the distribution of benthic photosynthetic organisms and their contribution to primary production. Biogeosciences 3, 489–513. https://doi.org/10.5194/bg-3-489-2006 (2006).Article 

Google Scholar 
Wright, D. H. Species-energy theory: An extension of species-area theory. Oikos 41, 496–506 (1983).Article 

Google Scholar 
Cusens, J., Wright, S. D., McBride, P. D. & Gillman, L. N. What is the form of the productivity–animal-species-richness relationship? A critical review and meta-analysis. Ecology 93, 2241–2252. https://doi.org/10.1890/11-1861.1 (2012).Article 
PubMed 

Google Scholar 
Rosenzweig, M. L. & Abramsky, Z. in Species Diversity in Ecological Communities. Historical and Geographical Perspectives (eds Ricklefs, R. E. & Schluter, D.) Ch. 5, 52–65 (The University of Chicago Press, 1993).Abrams, P. A. Monotonic or unimodal diversity-productivity gradients: What does competition theory predict?. Ecology 76, 2019–2027 (1995).Article 

Google Scholar 
Huston, M. A. Disturbance, productivity, and species diversity: Empiricism vs. logic in ecological theory. Ecology 95, 2382–2396 (2014).Article 

Google Scholar 
Roberts, T. E. et al. Testing biodiversity theory using species richness of reef-building corals across a depth gradient. Biol. Lett. 15, 20190493. https://doi.org/10.1098/rsbl.2019.0493 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Frankowiak, K. et al. Photosymbiosis and the expansion of shallow-water corals. Sci. Adv. 2, e1601122. https://doi.org/10.1126/sciadv.1601122 (2016).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Goreau, T. F. & Goreau, N. I. The physiology of skeleton formation in corals. II. Calcium deposition by hermatypic corals under various conditions in the reef. Biol. Bull. 117, 239–250. https://doi.org/10.2307/1538903 (1959).Article 
CAS 

Google Scholar 
Kirk, J. T. O. Light and Photosynthesis in Aquatic Ecosystems 3rd edn. (Cambridge University Press, 2011).
Google Scholar 
Stoddart, D. R. Ecology and morphology of recent coral reefs. Biol. Rev. 44, 433–498. https://doi.org/10.1111/j.1469-185X.1969.tb00609.x (1969).Article 

Google Scholar 
Lesser, M. P., Slattery, M. & Leichter, J. J. Ecology of mesophotic coral reefs. J. Exp. Mar. Biol. Ecol. 375, 1–8 (2009).Article 

Google Scholar 
Ackleson, S. G. Light in shallow waters: A brief research review. Limnol. Oceanogr. 48, 323–328. https://doi.org/10.4319/lo.2003.48.1_part_2.0323 (2003).Article 

Google Scholar 
Connell, J. H. Diversity in tropical rain forests and coral reefs. High diversity of trees and corals is maintained only in a nonequilibrium state. Science 199, 1302–1310. https://doi.org/10.1126/science.199.4335.1302 (1978).Article 
CAS 
PubMed 

Google Scholar 
Dollar, S. J. Wave stress and coral community structure in Hawaii. Coral Reefs 1, 71–81. https://doi.org/10.1007/BF00301688 (1982).Article 

Google Scholar 
Hughes, T. P. Community structure and diversity of coral reefs: The role of history. Ecology 70, 275–279. https://doi.org/10.2307/1938434 (1989).Article 

Google Scholar 
Fraser, R. H. & Currie, D. J. The species richness-energy hypothesis in a system where historical factors are thought to prevail: Coral reefs. Am. Nat. 148, 138–159 (1996).Article 

Google Scholar 
Cornell, H. V. & Karlson, R. H. Coral species richness: Ecological versus biogeographical influences. Coral Reefs 19, 37–49 (2000).Article 

Google Scholar 
Bellwood, D. R., Hughes, T., Connolly, S. & Tanner, J. Environmental and geometric constraints on Indo-Pacific coral reef biodiversity. Ecol. Lett. 8, 643–651. https://doi.org/10.1111/j.1461-0248.2005.00763.x (2005).Article 

Google Scholar 
Brown, B. E. et al. Diurnal changes in photochemical efficiency and xanthophyll concentrations in shallow water reef corals: Evidence for photoinhibition and photoprotection. Coral Reefs 18, 99–105 (1999).Article 

Google Scholar 
Hoegh-Guldberg, O. & Jones, R. J. Photoinhibition and photoprotection in symbiotic dinoflagellates from reef-building corals. Mar. Ecol. Prog. Ser. 183, 73–86. https://doi.org/10.3354/meps183073 (1999).Article 

Google Scholar 
Lesser, M. P. & Gorbunov, M. Y. Diurnal and bathymetric changes in chlorophyll fluorescence yields of reef corals measured in situ with a fast repetition rate fluorometer. Mar. Ecol. Prog. Ser. 212, 69–77. https://doi.org/10.3354/meps212069 (2001).Article 
CAS 

Google Scholar 
Hoogenboom, M. O., Anthony, K. R. N. & Connolly, S. R. Energetic cost of photoinhibition in corals. Mar. Ecol. Prog. Ser. 313, 1–12. https://doi.org/10.3354/meps313001 (2006).Article 
CAS 

Google Scholar 
Huot, Y. & Babin, M. Chlorophyll a Fluorescence in Aquatic Sciences: Methods and Applications 31–74 (Springer, 2010).Book 

Google Scholar 
Warner, M. E., Lesser, M. P. & Ralph, P. J. Chlorophyll a Fluorescence in Aquatic Sciences: Methods and Applications Ch. Chapter 10, 209–222 (Springer Science+Business Media B.V., 2010).Skirving, W. et al. Remote sensing of coral bleaching using temperature and light: Progress towards an operational algorithm. Remote Sens. 10, 18 (2018).Article 

Google Scholar 
Enríquez, S., Merino, M. & Iglesias-Prieto, R. Variations in the photosynthetic performance along the leaves of the tropical seagrass Thalassia testudinum. Mar. Biol. 140, 891–900. https://doi.org/10.1007/s00227-001-0760-y (2002).Article 
CAS 

Google Scholar 
Sundby, C., McCaffery, S. & Anderson, J. M. Turnover of the photosystem II D1 protein in higher plants under photoinhibitory and nonphotoinhibitory irradiance. J. Biol. Chem. 268, 25476–25482 (1993).Article 
CAS 
PubMed 

Google Scholar 
Tyystjärvi, E. & Aro, E. M. The rate constant of photoinhibition, measured in lincomycin-treated leaves, is directly proportional to light intensity. Proc. Natl. Acad. Sci. U. S. A. 93, 2213–2218. https://doi.org/10.1073/pnas.93.5.2213 (1996).Article 
PubMed 
PubMed Central 

Google Scholar 
Iglesias-Prieto, R., Beltrán, V. H., LaJeunesse, T. C., Reyes-Bonilla, H. & Thomé, P. E. Different algal symbionts explain the vertical distribution of dominant reef corals in the eastern Pacific. Proc. R. Soc. Lond. B 271, 1757–1763. https://doi.org/10.1098/rspb.2004.2757 (2004).Article 
CAS 

Google Scholar 
Jassby, A. D. & Platt, T. Mathematical formulation of the relationship between photosynthesis and light for phytoplankton. Limnol. Oceanogr. 21, 540–547 (1976).Article 
CAS 

Google Scholar 
Long, S. P., Humphries, S. & Falkowski, P. G. Photoinhibition of photosynthesis in nature. Annu. Rev. Plant Physiol. Plant Mol. Biol. 45, 633–662. https://doi.org/10.1146/annurev.pp.45.060194.003221 (1994).Article 
CAS 

Google Scholar 
Huner, N. P. A., Öuist, G. & Sarhan, F. Energy balance and acclimation to light and cold. Trends Plant Sci. 3, 224–230 (1998).Article 

Google Scholar 
Sheppard, C. R. C. Coral cover, zonation and diversity on reef slopes of Chagos Atolls, and population structures of the major species. Mar. Ecol. Prog. Ser. 2, 193–205 (1980).Article 

Google Scholar 
Huston, M. A. Patterns of species diversity in relation to depth at Discovery Bay, Jamaica. Bull. Mar. Sci. 37, 928–935 (1985).
Google Scholar 
Loya, Y. Community structure and species diversity of hermatypic corals at Eilat, Red Sea. Mar. Biol. 13, 100–123. https://doi.org/10.1007/BF00366561 (1972).Article 

Google Scholar 
Chow, G. S. E., Chan, Y. K. S., Jain, S. S. & Huang, D. Light limitation selects for depth generalists in urbanised reef coral communities. Mar. Environ. Res. 147, 101–112. https://doi.org/10.1016/j.marenvres.2019.04.010 (2019).Article 
CAS 
PubMed 

Google Scholar 
Kahng, S. E. et al. Community ecology of mesophotic coral reef ecosystems. Coral Reefs 29, 255–275. https://doi.org/10.1007/s00338-010-0593-6 (2010).Article 

Google Scholar 
Iglesias-Prieto, R. Temperature-dependent inactivation of Photosystem II in symbiotic dinoflagellates. in Proc. 8th Int. Coral Reef Sym, 1313–1318 (1997).Jones, R. J., Hoegh-Guldberg, O., Larkum, A. W. D. & Schreiber, U. Temperature-induced bleaching of corals begins with impairment of the CO2 fixation mechanism in zooxanthellae. Plant Cell Environ. 21, 1219–1230. https://doi.org/10.1046/j.1365-3040.1998.00345.x (1998).Article 
CAS 

Google Scholar 
Hennige, S. J., Suggett, D. J., Warner, M. E., McDougall, K. E. & Smith, D. J. Photobiology of Symbiodinium revisited: Bio-physical and bio-optical signatures. Coral Reefs 28, 179–195. https://doi.org/10.1007/s00338-008-0444-x (2008).Article 

Google Scholar 
Quigg, A. & Beardall, J. Protein turnover in relation to maintenance metabolism at low photon flux in two marine microalgae. Plant Cell Environ. 26, 693–703. https://doi.org/10.1046/j.1365-3040.2003.01004.x (2003).Article 
CAS 

Google Scholar 
Järvi, S., Suorsa, M. & Aro, E. M. Photosystem II repair in plant chloroplasts—Regulation, assisting proteins and shared components with photosystem II biogenesis. Biochim. Biophys. Acta Bioenerg. 900–909, 2015. https://doi.org/10.1016/j.bbabio.2015.01.006 (1847).Article 
CAS 

Google Scholar 
Jokiel, P. L. Solar ultraviolet radiation and coral reef epifauna. Science 207, 1069–1071 (1980).Article 
CAS 
PubMed 

Google Scholar 
López-Londoño, T. et al. Physiological and ecological consequences of the water optical properties degradation on reef corals. Coral Reefs 40, 1243–1256. https://doi.org/10.1007/s00338-021-02133-7 (2021).Article 

Google Scholar 
Vermeij, M. J. A. & Bak, R. P. M. How are coral populations structured by light? Marine light regimes and the distribution of Madracis. Mar. Ecol. Prog. Ser. 233, 105–116. https://doi.org/10.3354/meps233105 (2002).Article 

Google Scholar 
Hoogenboom, M. O., Connolly, S. R. & Anthony, K. R. N. Interactions between morphological and physiological plasticity optimize energy acquisition in corals. Ecology 89, 1144–1154. https://doi.org/10.1890/07-1272.1 (2008).Article 
PubMed 

Google Scholar 
Kaniewska, P., Anthony, K., Sampayo, E., Campbell, P. & Hoegh-Guldberg, O. Implications of geometric plasticity for maximizing photosynthesis in branching corals. Mar. Biol. 161, 313–328 (2014).Article 
CAS 

Google Scholar 
Kramer, N., Tamir, R., Eyal, G. & Loya, Y. Coral morphology portrays the spatial distribution and population size-structure along a 5–100 m depth gradient. Front. Mar. Sci. https://doi.org/10.3389/fmars.2020.00615 (2020).Article 

Google Scholar 
Lesser, M. P., Mobley, C. D., Hedley, J. D. & Slattery, M. Incident light on mesophotic corals is constrained by reef topography and colony morphology. Mar. Ecol. Prog. Ser. 670, 49–60. https://doi.org/10.3354/meps13756 (2021).Article 

Google Scholar 
Prada, C. et al. Linking photoacclimation responses and microbiome shifts between depth-segregated sibling species of reef corals. R. Soc. Open Sci. 9, 211591. https://doi.org/10.1098/rsos.211591 (2022).Article 
PubMed 
PubMed Central 

Google Scholar 
Rowan, R., Knowlton, N., Baker, A. & Jara, J. Landscape ecology of algal symbionts creates variation in episodes of coral bleaching. Nature 388, 265–269. https://doi.org/10.1038/40843 (1997).Article 
CAS 
PubMed 

Google Scholar 
Warner, M. E., LaJeunesse, T. C., Robison, J. D. & Thur, R. M. The ecological distribution and comparative photobiology of symbiotic dinoflagellates from reef corals in Belize: Potential implications for coral bleaching. Limnol. Oceanogr. 51, 1887–1897. https://doi.org/10.4319/lo.2006.51.4.1887 (2006).Article 

Google Scholar 
Anthony, K. R. N. & Fabricius, K. E. Shifting roles of heterotrophy and autotrophy in coral energetics under varying turbidity. J. Exp. Mar. Biol. Ecol. 252, 221–253 (2000).Article 
CAS 
PubMed 

Google Scholar 
Hoogenboom, M., Rodolfo-Metalpa, R. & Ferrier-Pagès, C. Co-variation between autotrophy and heterotrophy in the Mediterranean coral Cladocora caespitosa. J. Exp. Biol. 213, 2399–2409 (2010).Article 
PubMed 

Google Scholar 
Carlson, R. R., Foo, S. A. & Asner, G. P. Land use impacts on coral reef health: A ridge-to-reef perspective. Front. Mar. Sci 6, 562. https://doi.org/10.3389/fmars.2019.00562 (2019).Article 

Google Scholar 
Wang, M. et al. The great Atlantic Sargassum belt. Science 365, 83–87. https://doi.org/10.1126/science.aaw7912 (2019).Article 
CAS 
PubMed 

Google Scholar 
Alvarez-Filip, L., González-Barrios, F. J., Pérez-Cervantes, E., Molina-Hernández, A. & Estrada-Saldívar, N. Stony coral tissue loss disease decimated Caribbean coral populations and reshaped reef functionality. Commun. Biol. 5, 440. https://doi.org/10.1038/s42003-022-03398-6 (2022).Article 
PubMed 
PubMed Central 

Google Scholar 
Muscatine, L., McCloskey, L. R. & Marian, R. E. Estimating the daily contribution of carbon from zooxanthellae to coral animal respiration. Limnol. Oceanogr. 26, 601–611. https://doi.org/10.4319/lo.1981.26.4.0601 (1981).Article 
CAS 

Google Scholar 
Jørgensen, S. E. & Bendoricchio, G. Fundamentals of Ecological Modelling 3rd edn, Vol. 21 (Elsevier Sceince B. V., 2001).
Google Scholar 
Hennige, S. J. et al. Acclimation and adaptation of scleractinian coral communities along environmental gradients within an Indonesian reef system. J. Exp. Mar. Biol. Ecol. 391, 143–152. https://doi.org/10.1016/j.jembe.2010.06.019 (2010).Article 

Google Scholar 
Scheufen, T., Iglesias-Prieto, R. & Enríquez, S. Changes in the number of symbionts and Symbiodinium cell pigmentation modulate differentially coral light absorption and photosynthetic performance. Front. Mar. Sci 4, 309. https://doi.org/10.3389/fmars.2017.00309 (2017).Article 

Google Scholar 
Veron, J. E. N. Corals in Space and Time. The Biogeography and Evolution of the Scleractinia 321 (Cornell University Press, 1995).
Google Scholar 
Nelder, J. A. & Mead, R. A simplex method for function minimization. J. Comput. 7, 308–313. https://doi.org/10.1093/comjnl/7.4.308 (1965).Article 
MathSciNet 
MATH 

Google Scholar 
R: A languate and environment for statistical computing. Retrieved from http://www.R-project.org (R Foundation for Statistical Computing, Vienna, Austria, 2010). More

Versluys, J. Die Kaubewegungen von Trachodon. Palaontol. Z. 4, 80–87 (1922).
Google Scholar 
Kripp, D. Die Kaubewegung und Lebensweise von Edmontosaurus spec. auf Grund der mechanischkonstruktiven analyse. Palaeobiologica 5, 409–422 (1933).
Google Scholar 
Ostrom, J. H. Cranial morphology of the hadrosaurian dinosaurs of North America. Bull. Am. Mus. Nat. Hist. 122, 39–186 (1961).
Google Scholar 
Ostrom, J. H. A functional analysis of jaw mechanics in the dinosaur. Triceratops. Postilla. 88, 1–35 (1964).MathSciNet 

Google Scholar 
Galton, P. M. The cheeks of ornithischian dinosaurs. Lethaia 6, 67–89. https://doi.org/10.1111/j.1502-3931.1973.tb00873.x (1973).Article 

Google Scholar 
Galton, P. M. Herbivorous adaptations of Late Triassic and Early Jurassic dinosaurs. In The Beginning of the Age of Dinosaurs (ed. Padian, K.) 203–221 (Cambridge University Press, 1986).
Google Scholar 
Weishampel, D. B. Hadrosaurid jaw mechanics. Acta Palaeontol. Pol. 28, 271–280 (1983).
Google Scholar 
Weishampel, D. B. The evolution of jaw mechanisms in ornithopod dinosaurs. Adv. Anat. Embryol. Cell. Biol. 87, 1–2 (1984).Article 
CAS 
PubMed 

Google Scholar 
Weishampel, D. B. Interactions between Mesozoic plants and vertebrates: fructifications and seed predation. Neues Jahrb. Geol. Paläontol. Abh. 167, 224–250 (1984).
Google Scholar 
Weishampel, D. B. & Norman, D. B. Vertebrate herbivory in the Mesozoic: Jaws, plants, and evolutionary metrics. In Paleobiology of the Dinosaurs Special Papers 238 (ed. Farlow, J. O.) 87–100 (Geological Society of America, 1989).Chapter 

Google Scholar 
Norman, D. B. & Weishampel, D. B. Feeding mechanisms in some small herbivorous dinosaurs: processes and patterns. In Biomechanics and Evolution (eds Rayner, J. M. V. & Wooton, R. J.) 161–181 (Cambridge University Press, 1991).
Google Scholar 
Sereno, P., Zijin, Z. & Lin, T. A new psittacosaur from Inner Mongolia and the parrot-like structure and function of the psittacosaur skull. Proc. Roy. Soc. B. 277, 199–209. https://doi.org/10.1098/rspb.2009.0691 (2010).Article 

Google Scholar 
Barrett, P. M. Paleobiology of herbivorous dinosaurs. Annu. Rev. Earth Planet. Sci. 42(1), 207–230. https://doi.org/10.1146/annurev-earth-042711-105515 (2014).Article 
CAS 

Google Scholar 
Erickson, G. M. et al. Wear biomechanics in the slicing dentition of the giant horned dinosaur Triceratops. Sci. Adv. 1(5), e1500055. https://doi.org/10.1126/sciadv.1500055 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Nabavizadeh, A. Hadrosauroid jaw mechanics and the functionalsignificance of the predentary bone. In The hadrosaurs: Proceedings of the International Hadrosaur Symposium (eds Evans, D. & Eberth, D.) 467–482 (Indiana University Press, 2014).
Google Scholar 
Nabavizadeh, A. Evolutionary trends in the jaw adductor mechanics of ornithischian dinosaurs. Anat. Rec. 299(3), 271–294. https://doi.org/10.1002/ar.23306 (2016).Article 

Google Scholar 
Nabavizadeh, A. new reconstruction of cranial musculature in ceratopsian dinosaurs: Implications for jaw mechanics and ‘cheek’anatomy. FASEB J. 30, lb27–lb27. https://doi.org/10.1096/fasebj.30.1_supplement.lb27 (2016).Article 

Google Scholar 
Nabavizadeh, A. new reconstruction of cranial musculature in ornithischian dinosaurs: Implications for feeding mechanismsand buccal anatomy. Anat. Rec. 303, 347–362. https://doi.org/10.1002/ar.23988 (2020).Article 

Google Scholar 
Varriale, F. J. Dental microwear reveals mammal-like chewing in the neoceratopsian dinosaur Leptoceratops gracilis. PeerJ 4, e2132. https://doi.org/10.7717/peerj.2132 (2016).Article 
PubMed 
PubMed Central 

Google Scholar 
Melstrom, K. M., Chiappe, L. M. & Smith, N. D. Exceptionally simple, rapidly replaced teeth in sauropod dinosaurs demonstrate a novel evolutionary strategy for herbivory in Late Jurassic ecosystems. BMC Evol. Biol. 21(1), 1–12. https://doi.org/10.1186/s12862-021-01932-4 (2021).Article 

Google Scholar 
Norman, D. B. On the cranial morphology and evolution of ornithopod dinosaurs. Proc. Zool. Soc. Lond. 52, 521–547 (1984).
Google Scholar 
Norman, D. B. & Weishampel, D. B. Ornithopod feeding mechanisms: Their bearing on the evolution of herbivory. Am. Nat. 126, 151–164. https://doi.org/10.1086/284406 (1985).Article 

Google Scholar 
Norman, D. B. & Weishampel, D. B. Vegetarian dinosaurs chew it differently-living mammals can chew plants for more effectively than reptiles. Yet some dinosaurs were surprisingly adept chewers. This unexpected ability may have been crucial in their evolution. New Sci. 114(1559), 42–45 (1987).
Google Scholar 
Rybczynski, N., Tirabasso, A., Bloskie, P., Cuthbertson, R. & Holliday, C. A three-dimensional animation model of Edmontosaurus (Hadrosauridae) for testing chewing hypotheses. Palaeontol. Electron. 11(2), 9A (2008).
Google Scholar 
Williams, V. S., Barrett, P. M. & Purnell, M. A. Quantitative analysis of dental microwear in hadrosaurid dinosaurs, and the implications for hypotheses of jaw mechanics and feeding. PNAS 106(27), 11194–11199. https://doi.org/10.1073/pnas.0812631106 (2009).Article 
PubMed 
PubMed Central 

Google Scholar 
Cuthbertson, R. S., Tirabasso, A., Rybczynski, N. & Holmes, R. B. Kinetic limitations of intracranial joints in Brachylophosaurus canadensis and Edmontosaurus regalis (Dinosauria: Hadrosauridae), and their implications for the chewing mechanics of hadrosaurids. Anat. Rec. 295, 968–979. https://doi.org/10.1002/ar.22458 (2012).Article 

Google Scholar 
Erickson, G. M. & Zelenitsky, D. K. Osteohistology and occlusal morphology of Hypacrosaurus stebengeri teeth throughout ontogeny with comments on wear-induced form and function. In Hadrosaurs (eds Eberth, D. A. & Evans, D. C.) 422–432 (Indiana University Press, 2014).
Google Scholar 
Barrett, P. M. Tooth wear and possible jaw action of Scelidosaurus harrisonii Owen and a review of feeding mechanisms in other thyreophoran dinosaurs. In The Armored Dinosaurs (ed. Carpenter, K.) 25–52 (Indiana University Press, 2001).
Google Scholar 
Rybczynski, N. & Vickaryous, M. K. Evidence of complex jaw movement in the Late Cretaceous ankylosaurid Euoplocephalus tutus (Dinosauria: Thyreophora). In The Armored Dinosaurs (ed. Carpenter, K.) 299–317 (Indiana University Press, 2001).
Google Scholar 
Mallon, J. C. & Anderson, J. S. The functional and palaeoecological implications of tooth morphology and wear for the megaherbivorous dinosaurs from the Dinosaur Park Formation (Upper Campanian) of Alberta, Canada. PLoS ONE 9(6), e98605. https://doi.org/10.1371/journal.pone.0098605 (2014).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Mallon, J. C. & Anderson, J. S. Implications of beak morphology for the evolutionary paleoecology of the megaherbivorous dinosaurs from the Dinosaur Park Formation (upper Campanian) of Alberta, Canada. Palaeogeogr. Palaeoclimatol. Palaeoecol. 394, 29–41. https://doi.org/10.1016/j.palaeo.2013.11.014 (2014).Article 

Google Scholar 
Ősi, A., Barrett, P. M., Földes, T. & Tokai, R. Wear pattern, dental function, and jaw mechanism in the Late Cretaceous ankylosaur Hungarosaurus. Anat. Rec. 297(7), 1165–1180. https://doi.org/10.1002/ar.22910 (2014).Article 

Google Scholar 
Ősi, A., Prondvai, E., Mallon, J. & Bodor, E. R. Diversity and convergences in the evolution of feeding adaptations in ankylosaurs (Dinosauria: Ornithischia). Hist. Biol. 29(4), 539–570. https://doi.org/10.1080/08912963.2016.1208194 (2017).Article 

Google Scholar 
Hill, R. V., D’Emic, M. D., Bever, G. S. & Norell, M. A. A complex hyobranchial apparatus in a Cretaceous dinosaur and the antiquity of avian paraglossalia. Zool. J. Linn. Soc. 175(4), 892–909. https://doi.org/10.1111/zoj.12293 (2015).Article 

Google Scholar 
Lautenschlager, S., Brassey, C. A., Button, D. J. & Barrett, P. M. Decoupled form and function in disparate herbivorous dinosaur clades. Sci. Rep. 6(1), 1–10. https://doi.org/10.1038/srep26495 (2016).Article 
CAS 

Google Scholar 
Skutschas, P. P. et al. Wear patterns and dental functioning in an Early Cretaceous stegosaur from Yakutia, Eastern Russia. PLoS ONE 16(3), e0248163. https://doi.org/10.1371/journal.pone.0248163 (2021).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Strickson, E., Prieto-Márquez, A., Benton, M. J. & Stubbs, T. L. Dynamics of dental evolution in ornithopod dinosaurs. Sci. Rep. 6, 28904. https://doi.org/10.1038/srep28904 (2016).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Virág, A. & Ősi, A. Morphometry, microstructure, and wear pattern of neornithischian dinosaur teeth from the Upper Cretaceous Iharkút locality (Hungary). Anat. Rec. 300(8), 1439–1463. https://doi.org/10.1002/ar.23592 (2017).Article 

Google Scholar 
Mallon, J. C. & Anderson, J. S. Skull ecomorphology of megaherbivorous dinosaurs from the Dinosaur Park Formation (Upper Campanian) of Alberta, Canada. PLoS ONE 8(7), e67182. https://doi.org/10.1371/journal.pone.0067182 (2013).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Botfalvai, G., Ősi, A. & Mindszenty, A. Taphonomic and paleoecologic investigations of the Late Cretaceous (Santonian) Iharkút vertebrate assemblage (Bakony Mts, northwestern Hungary). Palaeogeogr. Palaeoclimatol. Palaeoecol. 417, 379–405. https://doi.org/10.1016/j.palaeo.2014.09.032 (2015).Article 

Google Scholar 
Botfalvai, G., Haas, J., Bodor, E. R., Mindszenty, A. & Ősi, A. Facies architecture and palaeoenvironmental implications of the upper Cretaceous (Santonian) Csehbánya formation at the Iharkút vertebrate locality (Bakony Mountains, Northwestern Hungary). Palaeogeogr. Palaeoclimatol. Palaeoecol. 441, 659–678. https://doi.org/10.1016/j.palaeo.2015.10.018 (2016).Article 

Google Scholar 
Ősi, A. et al. The Late Cretaceous continental vertebrate fauna from Iharkút, western Hungary: A review. In Bernissart Dinosaurs and Early Cretaceous Terrestrial Ecosystems (ed. Godefroit, P.) 532–569 (Indiana University Press, 2012).
Google Scholar 
Wells, N. A. Making thin sections. In Paleotechniques (eds Feldmann, R. M. et al.) 120–129 (University of Tennessee, 1989).
Google Scholar 
Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9(7), 671–675. https://doi.org/10.1038/nmeth.2089 (2012).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Evans, A. R. Surfer Manipulator. http://evomorph.org/surfermanipulator (2011).Evans, A. R., Wilson, G. P., Fortelius, M. & Jernvall, J. High-level similarity of dentitions in carnivorans and rodents. Nature 445, 78–81. https://doi.org/10.1038/nature05433 (2007).Article 
CAS 
PubMed 

Google Scholar 
Wilson, G. P. et al. Adaptive radiation of multituberculate mammals before the extinction of dinosaurs. Nature 483, 457–460. https://doi.org/10.1038/nature10880 (2012).Article 
CAS 
PubMed 

Google Scholar 
Ungar, P. S. Dental microwear of European Miocene catarrhines: Evidence for diets and tooth use. J. Hum. Evol. 31, 355–366. https://doi.org/10.1006/jhev.1996.0065 (1996).Article 

Google Scholar 
Ungar, P. S. A semiautomated image analysis procedure for the quantification of dental microwear II. Scanning. 17, 57–59. https://doi.org/10.1002/sca.4950170108 (1995).Article 
CAS 
PubMed 

Google Scholar 
Ungar, P. S., Brown, C. A., Bergstrom, T. S. & Walker, A. Quantification of dental microwear by tandem scanning confocal microscopy and scale-sSensitive fractal analyses. Scanning 25, 185–193. https://doi.org/10.1002/sca.4950250405 (2003).Article 
PubMed 

Google Scholar 
Ungar, P. S., Merceron, G. & Scott, R. S. Dental microwear texture analysis of Varswater bovids and Early Pliocene paleoenvironments of langebaanweg, Western Cape Province, South Africa. J. Mammal. Evol. 14, 163–181. https://doi.org/10.1007/s10914-007-9050-x (2007).Article 

Google Scholar 
Scott, J. R. Dental microwear texture analysis of extant African Bovidae. Mammalia 76, 157–217. https://doi.org/10.1515/mammalia-2011-0083 (2012).Article 

Google Scholar 
Merceron, G., Hofman-Kaminska, E. & Kowalczyk, R. 3D dental microwear texture analysis of feeding habits of sympatric ruminants in the Białowieza Primeval Forest, Poland. For. Ecol. Manag. 328, 262–269. https://doi.org/10.1016/j.foreco.2014.05.041 (2014).Article 

Google Scholar 
Caporale, S. S. & Ungar, P. S. Rodent incisor microwear as a proxy for ecological reconstruction. Palaeogeog. Palaeocl. Palaeoecol. 446, 225–233. https://doi.org/10.1016/j.palaeo.2016.01.013 (2016).Article 

Google Scholar 
R Core Team. R. A language and environment for statistical computing. R Foundation for Statistical Computing https://www.R-project.org/ (2021).Erickson, G. M. Incremental lines of von Ebner in dinosaurs and the assessment of tooth replacement rates using growth line counts. PNAS 93(25), 14623–14627. https://doi.org/10.1073/pnas.93.25.14623 (1996).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Godefroit, P. et al. Extreme tooth enlargement in a new Late Cretaceous rhabdodontid dinosaur from Southern France. Sci. Rep. 7(1), 1–9. https://doi.org/10.1038/s41598-017-13160-2 (2017).Article 
CAS 

Google Scholar 
Edmund, G. Tooth replacement phenomena in the lower vertebrates. Life. Sci. Contrib. R. Ont. Mus. 52, 1–190 (1960).
Google Scholar 
D’Emic, M. D., Whitlock, J. A., Smith, K. M., Fisher, D. C. & Wilson, J. A. Evolution of high tooth replacement rates in sauropod dinosaurs. PLoS ONE 8(7), e69235. https://doi.org/10.1371/journal.pone.0069235 (2013).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Ősi, A., Prondvai, E., Butler, R. & Weishampel, D. B. Phylogeny, histology and inferred body size evolution in a new rhabdodontid dinosaur from the Late Cretaceous of Hungary. PLoS ONE 7(9), e44318. https://doi.org/10.1371/journal.pone.0044318 (2012).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Weishampel, D. B., Jianu, C. M., Csiki, Z. & Norman, D. B. Osteology and phylogeny of Zalmoxes (ng), an unusual euornithopod dinosaur from the latest Cretaceous of Romania. J. Syst. Palaeontol. 1(2), 65–123. https://doi.org/10.1017/S1477201903001032 (2003).Article 

Google Scholar 
Melstrom, K. M. The relationship between diet and tooth complexity in living dentigerous saurians. J. Morphol. 278, 500–522 (2017).Article 
PubMed 

Google Scholar 
LeBlanc, A. R. H., Reisz, R. R., Evans, D. C. & Bailleul, A. M. Ontogeny reveals function and evolution of the hadrosaurid dinosaur dental battery. BMC Evol. Biol. 16(1), 1–13. https://doi.org/10.1186/s12862-016-0721-1 (2016).Article 

Google Scholar 
Erickson, G. M. et al. Complex dental structure and wear biomechanics in hadrosaurid dinosaurs. Science 338(6103), 98–101. https://doi.org/10.1126/science.1224495 (2012).Article 
CAS 
PubMed 

Google Scholar 
Norman, D. B. & Weishampel, D. B. Iguanodontidae and related Ornithopoda. In The Dinosauria (eds Weishampel, D. B. et al.) 510–533 (University of California Press, 1990).
Google Scholar 
Hulke, J. W. An attempt at a complete osteology of Hypsilophodon foxii, a British Wealden dinosaur. Philos. Trans. R. Soc. Lond. 172, 1035–1062. https://doi.org/10.1098/rstl.1882.0025 (1882).Article 

Google Scholar 
Sternberg, C. H. Thescelosaurus edmontonensis, n. sp., and classification of the Hypsilophodontidae. J. Paleontol. 14, 481–494 (1940).
Google Scholar 
Galton, P. M. The ornithischian dinosaur Hypsilophodon from the Wealden of the Isle of Wight. Bull. Br. Mus. Nat. Hist. 25(1), 1–152 (1974).
Google Scholar 
Norman, D. B. On the anatomy of Iguanodon atherfieldensis (Ornithischia: Ornithopoda). Bull. Inst. Roy. Sci. Nat. Belgique 56, 281–372 (1986).
Google Scholar 
Norman, D. B. & Barrett, P. M. Ornithischian dinosaurs from the lower Cretaceous (Berriasian) of England. Spec. Pap. Palaeontol. 68, 161–190 (2002).
Google Scholar 
Kosch, J. C. & Zanno, L. E. Sampling impacts the assessment of tooth growth and replacement rates in archosaurs: Implications for paleontological studies. PeerJ 8, e9918. https://doi.org/10.7717/peerj.9918 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Janis, C. M. & Fortelius, M. On the means whereby mammals achieve increased functional durability of their dentitions with special reference to limiting factors. Biol. Rev. 63, 197–230. https://doi.org/10.1111/j.1469-185X.1988.tb00630.x (1988).Article 
CAS 
PubMed 

Google Scholar 
You, H., Ji, Q. & Li, D. Lanzhousaurus magnidens gen. et sp. nov. from Gansu Province, China: The largest-toothed herbivorous dinosaur in the world. Geol. Bull. Chi 24(9), 785–794 (2005).
Google Scholar 
Suarez, C. A., You, H. L., Suarez, M. B., Li, D. Q. & Trieschmann, J. B. Stable isotopes reveal rapid enamel elongation (amelogenesis) rates for the Early Cretaceous iguanodontian dinosaur Lanzhousaurus magnidens. Sci. Rep. 7, 15319. https://doi.org/10.1038/s41598-017-15653-6 (2017).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Upchurch, P. & Barrett, P. M. The evolution of sauropod feeding mechanisms. In Evolution of Herbivory in Terrestrial Vertebrates: Perspectives from the Fossil Record (ed. Sues, H. D.) 79–122 (Cambridge University Press, 2000).Chapter 

Google Scholar 
Sereno, P. C. & Wilson, J. A. Structure and evolution of a sauropod tooth battery in Curry. In The Sauropods: Evolution and Paleobiology (eds Rogers, K. A. & Wilson, J. A.) 157–177 (University of California Press, 2005).
Google Scholar 
Brown, B. & Schlaikjer, E. M. The structure and relationships of Protoceratops. Ann. N. Y. Acad. Sci. 40(3), 133–265. https://doi.org/10.1111/j.1749-6632.1940.tb57047.x (1940).Article 

Google Scholar 
Solounias, N., Teaford, M. & Walker, A. Interpreting the diet of extinct ruminants-the case of a non-browsing giraffid. Paleobiology 14, 287–300. https://doi.org/10.1017/S009483730001201X (1988).Article 

Google Scholar 
Walker, A. & Teaford, M. Inferences from quantitative analysis of dental microwear. Folia Primatol. 53, 177–189. https://doi.org/10.1159/000156415 (1989).Article 
CAS 

Google Scholar 
Ungar, P. S. Mammalian dental function and wear: A review. Biosurf. Biotribol. 1(1), 25–41. https://doi.org/10.1016/j.bsbt.2014.12.001 (2015).Article 
MathSciNet 

Google Scholar 
Janis, C. M. An estimation of tooth volume and hypsodonty indices in ungulate mammals, and the correlation of these factors with dietary preferences. Mém. Mus. Natl. Hist. Nat. Sér. C Géol. 53, 367–387 (1988).
Google Scholar 
Lucas, P. W. et al. The role of dust, grit and phytoliths in tooth wear. Ann. Zool. Fenn. 51(1–2), 143–152. https://doi.org/10.5735/086.051.0215 (2014).Article 

Google Scholar 
Winkler, D. E. et al. Shape, size, and quantity of ingested external abrasives influence dental microwear texture formation in guinea pigs. Proc. Nat. Acad. Sci. 117, 22264–22273. https://doi.org/10.1073/pnas.2008149117 (2020).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Kaiser, T. M. et al. Nano-indentation of native phytoliths and dental tissues: Implications for herbivore-plant combat and dental wear proxies. Evol. Syst. 2, 55–63. https://doi.org/10.3897/evolsyst.2.22678 (2018).Article 

Google Scholar 
Winkler, D. E. et al. Forage silica and water content control dental surface texture in guinea pigs and provide implications for dietary reconstruction. Proc. Nat. Acad. Sci. 116, 1325–1330. https://doi.org/10.1073/pnas.1814081116 (2019).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Ősi, A. & Makádi, L. New remains of Hungarosaurus tormai (Ankylosauria, Dinosauria) from the Upper Cretaceous of Hungary: Skeletal reconstruction and body mass estimation. Palaontol. Z. 83(2), 227–245. https://doi.org/10.1007/s12542-009-0017-5 (2009).Article 

Google Scholar 
Winkler, D. E., Schulz-Kornas, E., Kaiser, T. M. & Tütken, T. Dental microwear texture reflects dietary tendencies in extant Lepidosauria despite their limited use of oral food processing. Proc. R. Soc. B 286, 20190544. https://doi.org/10.1098/rspb.2019.0544 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Bestwick, J., Unwin, D. M., Butler, R. J. & Purnell, M. A. Dietary diversity and evolution of the earliest flying vertebrates revealed by dental microwear texture analysis. Nat. Commun. 11, 1–9. https://doi.org/10.1038/s41467-020-19022-2 (2020).Article 
CAS 

Google Scholar 
Sakaki, H. et al. Non-occlusal dental microwear texture analysis of a titanosauriform sauropod dinosaur from the Upper Cretaceous (Turonian) Tamagawa Formation, northeastern Japan. Cret. Res. 136, 105218. https://doi.org/10.1016/j.cretres.2022.105218 (2022).Article 

Google Scholar 
Fiorillo, A. R. Dental microwear on the teeth of Camarasaurus and Diplodocus; implications for sauropod paleoecology. In Fifth Symposium on Mesozoic Terrestrial Ecosystems and Biota (eds Kielan-Jaworowska, Z. et al.) 23–24 (Paleontologisk Museum, 1991).
Google Scholar 
Mallon, J. C., Cuthbertson, R. S. & Tirabasso, A. Hadrosaurid jaw mechanics as revealed by cranial joint limitations and dental microwear analysis. In Hadrosaur Symposium Abstract Volume (eds Braman, D. R. et al.) 87–90 (Royal Tyrrell Museum of Palaeontology, 2011).
Google Scholar 
Fiorillo, A. R. Dental microwear patterns of the sauropod dinosaurs Camarasaurus and Diplodocus: Evidence for resource partitioning in the Late Jurassic of North America. Hist. Biol. 13, 1–16. https://doi.org/10.1080/08912969809386568 (1998).Article 

Google Scholar 
Sereno, P. C. et al. Structural extremes in a Cretaceous dinosaur. PLoS ONE 2(11), e1230. https://doi.org/10.1371/journal.pone.0001230 (2007).Article 
PubMed 
PubMed Central 

Google Scholar 
Whitlock, J. A. Inferences of diplodocoid (Sauropoda: Dinosauria) feeding behavior from snout shape and microwear analyses. PLoS ONE 6(4), e18304. https://doi.org/10.1371/journal.pone.0018304 (2011).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Fiorillo, A. R. Microwear patterns on the teeth of northern high latitude hadrosaurs with comments on microwear patterns in hadrosaurs as a function of latitude and seasonal ecological constraints. Palaeontol. Electron. 14(3), 20A (2011).
Google Scholar 
Bell, P. R., Snively, E. & Shychoski, L. A comparison of the jaw mechanics in hadrosaurid and ceratopsid dinosaurs using finite element analysis. Anat. Rec. 292(9), 1338–1351. https://doi.org/10.1002/ar.20978 (2009).Article 

Google Scholar 
Chin, K. & Gill, B. D. Dinosaurs, dung beetles, and conifers: Participants in a Cretaceous food web. Palaios 11, 280–285. https://doi.org/10.2307/3515235 (1996).Article 

Google Scholar 
Brown, C. M. et al. Dietary palaeoecology of an early Cretaceous armoured dinosaur (Ornithischia; Nodosauridae) based on floral analysis of stomach contents. Roy. Soc. Open Sci. 7(6), 200305. https://doi.org/10.1098/rsos.200305 (2020).Article 
CAS 

Google Scholar 
Crane, P. C., Friis, E. M. & Pedersen, K. R. The origin and early diversification of angiosperms. Nature 374, 27–33 (1995).Article 
CAS 

Google Scholar 
Friis, E. M., Crane, P. R. & Pedersen, K. R. Early Flowers and Angiosperm Evolution 1–596 (Cambridge University Press, 2011). https://doi.org/10.1017/CBO9780511980206.Book 

Google Scholar 
Benson, R. B., Hunt, G., Carrano, M. T. & Campione, N. Cope’s rule and the adaptive landscape of dinosaur body size evolution. Palaeontology 61, 13–48. https://doi.org/10.1111/pala.12329 (2018).Article 

Google Scholar 
Hummel, J. et al. In vitro digestibility of fern and gymnosperm foliage: Implications for sauropod feeing ecology and diet selection. Proc. Royal Soc. B 275, 1015–1021. https://doi.org/10.1098/rspb.2007.1728 (2008).Article 

Google Scholar 
Gee, C. T. Dietary options for the sauropod dinosaurs from an integrated botanical and paleobotanical perspective. In Biology of the Sauropod Dinosaurs: Understanding the Life of Giants (eds Klein, K. et al.) 34–56 (Indiana University Press, 2011).
Google Scholar 
Peters, R. H. The Ecological Implications of Body Size 1–329 (Cambridge University Press, 1983).Book 

Google Scholar 
Jarman, P. J. The social organisation of antelope in relation to their ecology. Behaviour 48, 215–267 (1974).Article 

Google Scholar  More