Philippa Kaur

Seed sampling and germinationSeed from ten trees from each of 21 native Scottish Scots pine populations (Table 1) were collected in March 2007 and germinated at the James Hutton Institute, Aberdeen (latitude 57.133214, longitude −2.158764) in June 2007. Populations were chosen to represent the species’ native range in Scotland and to include three populations from each of the seven seed zones (Fig. 2). There was no selection of seed-trees on the basis of any traits except for the possession of cones on the date of sampling. Ten seed trees were sampled from each population according to a spatial protocol designed to cover a circle of approximately 1 km in diameter located around a central tree. The sampling strategy identified nine points each in a pre-determined random direction from the central point, whilst stratifying the number sampled with increasing distance from the central point in the ratio 1: 3: 5. This strategy avoids over-sampling the areas close to the centre point. For smaller fragments of woodland, or where only a few trees with cones were present, then the directions of the sampled trees from the central tree were maintained to give a wide coverage of the woodland area, but the distances between trees varied but were never closer than 50 m. To break dormancy, seeds were soaked for 24 hours on the benchtop at room temperature, after which they were stored in wet paper towels and refrigerated in darkness at 3–5 °C for approximately 4 weeks. Seeds were kept moist and transferred to room temperature until germination began (approx. 5–7 days), then transplanted to 8 cm × 8 cm × 9 cm, 0.4 L pots filled with Levington’s C2a compost and 1.5 g of Osmocote Exact 16–18 months slow release fertiliser and kept in an unheated glasshouse. The compost was covered with a layer of grit to reduce moss and liverwort growth. Seedlings from the same mother tree are described as a family and are assumed to be half-siblings.Table 1 Locations and basic environmental data for the populations sampled for seed to establish the trial. See the maternal traits dataset15 for precise data for each mother tree sampled.Full size tableExperimental design: nurseriesThe full collection consisted of 210 families (10 families from each of 21 populations) each consisting of 24 half sibling progeny (total 5,040 individuals); needle tissue was sampled from each seedling and preserved for long term storage, one needle on silica gel, 2–5 needles at −20 °C. After transfer into pots, 8 seedlings per family were moved to one of three nurseries (total 1,680 seedlings per nursery): outdoors at Inverewe Gardens in western Scotland (nursery in the west of Scotland: coded NW, latitude 57.775714, longitude −5.597181, Fig. 2); outdoors in a fruit cage (to minimise browsing) at the James Hutton Institute in Aberdeen (nursery in the east of Scotland: NE); in an unheated glasshouse at the James Hutton Institute in Aberdeen (nursery in a glasshouse: NG). Trees were arranged in 40 randomised trays (blocks) in each nursery. Each block contained two trees per population (total 42 trees). Watering was automatic in NG, and manually as required for NE and NW. No artificial light was used in any of the nurseries. In May 2010 the seedlings from NG were moved outdoors to Glensaugh in Aberdeenshire (latitude 56.893567, longitude −2.535736). In 2010 all plants were repotted into 19 cm (3 L) pots containing Levingtons CNSE Ericaceous compost with added Osmocote STD 16–18 month slow release fertilizer.Experimental design: field sitesIn 2012 the trees were transplanted to one of three field sites: Yair in the Scottish Borders (field site in the south of Scotland: FS, latitude 55.603625, longitude −2.893025); Glensaugh (field site in the east of Scotland: FE); and Inverewe (field site in the west of Scotland: FW). All trees transplanted to FS were raised in the NG and all but four of the trees transplanted to FE were raised locally in the NE (the remainder were grown in NG). In contrast, following mortality and ‘beating up’ (filling gaps where saplings had died), the FW experiment ultimately contained cohorts of trees raised in each of the three nurseries as follows: 290 grown locally in the NW; 132 were grown in the NG; and 82 were grown in the NE.Site historiesThe Yair site (FS) had previously been used for growing Noble fir (Abies procera) for Christmas trees and Lodgepole pine (Pinus contorta), a section of the former were felled and chipped to create a clear area prior to planting. The planting site is also adjacent to a large block of commercial Sitka spruce (Picea sitchensis) forestry, and the Glenkinnon Burn Site of Special Scientific Interest (SSSI NatureScot site code 736; EU site code 135445), an area of mixed broadleaf woodland. Prior to planting, major areas of tall weeds were strimmed. The site was protected by a deer fence. The experiment was planted 8–11 October 2012. The Glensaugh site (FE) is in Forestry Compartment 3 of the Glensaugh Research Station, adjacent to Cleek Loch. It is thought to have been cleared of Scots pine and Larch (Larix decidua) around 1917, after which it reverted to rough grazing. An attempt to reseed part of the site in the 1980s was unsuccessful and it quickly reverted to rough grazing for a second time. The whole site (within which the experimental area is embedded) was deer fenced and re-planted under the Scottish Rural Development Programme (SRDP) in 2012. The experimental plot was planted up 7–9 March 2012. The Inverewe site (FW) had previously been a Sitka spruce and Lodgepole pine plantation (50:50 mix) that had been clear-felled in 2010 following substantial windthrow. The experimental site was deer fenced in early 2012, and the experiment was planted 12–16 March 2012, followed by beating up on 27–28 March 2013 and 22–24 October 2013. There had been minimal preparation of the site in line with current practice for restocking sites. The experimental site is included in the Inverewe Forest Plan, which included deer fencing of a larger area (2014) around the experimental site. Planting of this area was completed in early 2015, funded by NTS (National Trust for Scotland), although natural regeneration is also taking place.At each site, trees were planted in randomised blocks at 3 m × 3 m spacing. There are four randomised blocks in both FS and FE and three in FW. A guard row of Scots pine trees was planted around the periphery of the blocks and between blocks B and C at FS. Each block comprised one individual from each of eight (of the 10 sampled) families per 21 populations (168 trees). Although most families (N = 159) were represented at each of the three sites, families with insufficient trees (N = 9) were replaced in one site (FS) with a different family from the same population. Each experimental site was designed with redundancy such that, if thinning becomes necessary as the trees mature, then the systematic removal of trees (i.e. trees 1,3,5,7, etc of row 1, and 2,4,6,8, etc of row 2, 1,3,5,7,etc of row 3) will maintain a balanced design of the experiment, with sufficient family and population representation to provide an ongoing experiment with full geographic coverage.The field sites generally experience different climates, with FW typically warmer and wetter and with more growing degree days per year and a much longer growing season than both FE and FS (Table 2). The coldest site with the shortest growing season is generally FE.Table 2 Average climatic variables at field sites Glensaugh (FE), Inverewe (FW) and Yair (FS) from planting in 2012 until 2020. Climatic variables are derived from data provided by the Met Office (daily mean, minimum and maximum temperatures and monthly rainfall).Full size tablePhenotype assessmentsMaternal traitsFollowing seed collection, a range of traits were measured in the mother trees in order to control for maternal effects in subsequent measurements of their progeny (Table 3). For each mother tree, measurements of height and diameter at breast height (DBH) were taken, and ten cones were collected and assessed in detail. Cone width and length were measured prior to drying the cones (when they were still closed). Cone weight was measured post-drying. Seed removed from each cone was assessed for total weight (after wings had been removed) and for the count and percentage of seeds which were classed as viable (viable seed were those which had both a wing and an obvious seed). No further seed sorting was applied.Table 3 Traits assessed in mother trees, cones, seeds (dataset: Maternal), nursery seedlings (dataset: Nursery) and field trials (dataset: Field). Within the datasets, traits are recorded in a single column for each year using the format Code-Year (e.g. absolute height in 2008 = HA08) except for the maternal traits datasets which were all measured in 2007.Full size tableNursery traitsSeedling phenotype assessments were performed annually from 2007–2010 for three different trait types: phenology (budburst and growth cessation); form (total number of buds, needle length); cumulative growth (stem diameter and height, canopy width). Measurements of tree form and cumulative growth traits were taken after the end of each growing season. Phenology was assessed weekly during the spring and autumn of 2008 for budburst and growth cessation, respectively. Budburst was defined as the number of days from 31 March 2008 to the time when newly emerged green needles were observed (budburst stage 6: Fig. 3). In some seedlings in 2008, a secondary flush of growth occurred from terminal buds that had formed during the summer of that year. Therefore, growth cessation was defined retrospectively as the number of days from 10 September 2008 to the date when a terminal bud had formed on the leading shoot of the seedling, providing no further growth was observed either on the stem below that bud, nor from the bud itself. Canopy width (widest point) was measured at two perpendicular points in the horizontal plane. Needle length was measured for three needles per tree. Mortality was recorded each year from 2007 to 2010.Fig. 3Phenological stages of bud burst in Pinus sylvestris assessed in field trials. Inset numbers correspond to budburst stage. Budburst stage 1: bud dormant; 2: bud swelling and showing signs of linear expansion; 3: scales open at base revealing green tissue. Remaining bud remains encased by smooth bud scales; 4: scales open along length of shoot revealing green tissue and partially visible needles; 5: white tipped needles visible along length of the shoot; 6: green needles emerging away from the shoot (bottle brush appearance) along its entire length; 7: Needles have separated and next year’s terminal bud is usually formed and clearly visible.Full size imageField traitsTree height was measured in the field in the winter after each growing season from 2013 at FE and FW, and from 2014 to 2020 at all sites. Height was taken as the vertical measurement in cm from top bud straight to the ground. Basal stem diameter was measured at the end of the growing season for trees growing at FE and FW from 2014 to 2020 and for FS in 2020.Phenology assessments were performed in spring at each site from 2015 to 2019. Seven distinct stages of budburst (assessed on the terminal bud) were defined (Fig. 3) although only stages 4 to 6 are included in the dataset and considered for analysis due to high proportions of missing data for the early and late stages. Each tree was assessed for budburst stage at weekly intervals from early spring until budburst was complete. In order to allow comparisons within and among sites and years, the date at which each stage of budburst occurred was considered relative to 31 March of that year. For example, 25 May 2019 is recorded as 55 days since 31 March 2019. The duration of budburst (time taken to reach stage 6 from stage 4) was also estimated.When trees progressed through budburst stages rapidly, skipping a stage between assessments, a mean value was taken from the two assessment dates. For example, if a tree was at stage 4 on day 55 and was recorded as stage 6 at the next assessment on day 62, it is assumed to have reached stage 5 at day 58.5. More

SamplesThe intestine and muscle samples from 22 wild boars were collected between September 4 and March 2, 2020, in Namie town in Fukushima prefecture. Furthermore, control intestine samples were collected from three wild boars in Hyogo prefecture. Each location is depicted in Fig. 1. In each case, after the licensed hunters slaughtered the wild boar to be exterminated, only the tissue was transferred to the study.Measurement of radioactivityRadioactivity in the muscle samples was determined by gamma-ray spectrometry using high-purity germanium (HPGe) detectors (Ortec Co., Oak Ridge, TN, USA), as described in our previous report3. Gamma rays from 137Cs were observed.Exposure dose estimationIn order to estimate internal and external dose rates of the wild boars according to the ICRP publication 10826, we supposed the shapes of wild boars as prolate spheroids whose long axis was to be their body lengths. The short axis was given from their weight assuming their specific gravities were the same
as water. The dose rates were calculated from the contribution of 137Cs, not including
natural radionuclides. The energy deposition to the spheroids by beta and gamma rays from radionuclides were calculated by the numerical simulation with the use of the Particle and Heavy Ion Transport code System (PHITS)27. For the sake of simplicity, we supposed the spheroids consisted of only muscle, which would give overestimated values because muscle contains more radio cesium than other organs. The external exposure dose was calculated from the air dose rates which were observed from the monitoring post near the boars captured place. The average values of the air dose rates were obtained from fitting observed data of two years with decay curve. The background due to the natural radionuclides was estimated to be 0.05 µGy/h which was observed before the Fukushima Daiichi accident, and was removed before the fittings. The half-lives of the air dose rates were 2000–3000 days depending on the environment. Assuming the external exposure dose was ascribed to the 137Cs included in the surface of the ground. The amount of the 137Cs was calculated so as to reproduce the observed air does rates. Since the maximum range of the beta ray from 137Cs is a few millimeters, almost all of the beta ray from inside the body should be absorbed in the boar’s body, but the beta ray from outside the body would stop in its fur. The beta rays contribute 100% to internal exposure dose but 0% to external one. Since the linear attenuation coefficient for gamma rays from 137Cs is 0.084 cm−1 = (12 cm)−1, some of the gamma rays cannot stop in the body depending on the size of the body. The numerical simulation suggested that 65–90 percent of the gamma rays from 137Cs inside the body would go out, and 40–65 percent of the gamma rays from 137Cs outside would go through the body.Pathological analysisA piece of the small intestine was fixed in 10% neutral formalin at 4 °C for 24–48 h. Then, paraffin blocks were prepared for pathomorphological examination using hematoxylin and eosin (HE) staining.Gene expression analysisTotal RNA was extracted from the whole tissue of the intestine using TRIzol Reagent (Life Technologies, Inc., Frederic, MD, USA) according to the manufacturer’s instructions. RNA concentration was measured using a NanoDrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA), and cDNA was synthesized using random primers and SuperScript II (Life Technologies, Inc.). Real-time PCR for IFN-γ, TLR3, and CyclinG1 was performed using Brilliant SYBR Green QPCR Master Mix III (Stratagene, La Jolla, CA, USA) with an AriaMx system (Agilent Technologies, Santa Clara, CA, USA). Primer sequences were designed using Primer-BLAST with sequences obtained from GenBank as described in the previous report4. Amplification conditions were 95 °C for 3 min, 40 cycles at 95 °C for 5 s, and 60 °C for 20 s. Fluorescence signals measured during the amplification were analyzed. Ribosomal RNA primers were used as an internal control, and all data were normalized to constitutive rRNA values. Quantitative differences between the groups were analyzed using the AriaMx software (Agilent Technologies).Statistical analysisAll data are presented as mean ± standard error (SE) for each treatment group. Differences in mRNA expression among the groups were determined using the unpaired t-test with Welch’s correction. (Prism: GraphPad Software Inc., La Jolla, CA, USA). Differences were considered to be statistically significant at a P value of  More

The framework of Swin MSIWe established the first Swin Transformer-based mosquito species identification (Swin MSI) model, with the help of self-constructed image dataset and multi-adjustment-test. Gradient-weighted class activation mapping was used to visualize the identification process (Fig. 1a). The key Swin Transformer block was described on Fig. 1b. Based on practical needs, Swin MSI was additional designed to identify Culex pipiens Complex on the subspecies level (Fig. 1c) and novel mosquito (which was defined as ones beyond 17 species in our dataset) classification attribution (Fig. 1d). Detailed results are shown in the following sections.Figure 1The Framework of Swin MSI. (a)The basic architecture for mosquito features extraction and identification. Attention visualization generated by filters at each layer are shown. (b) Details for Swin Transformer block. (c) For mosquito within our dataset 17 species, output is the top 5 confidence species. (d) For mosquito beyond 17 species (defined as novel species), whether the output is a species or a genus is decided after comparing with confidence threshold.Full size imageMosquito datasetsWe established the highest-definition and most-balanced mosquito image dataset to date. The mosquito image dataset covers 7 genera and 17 species (including 3 morphologically similar subspecies in the Cx. pipiens Complex), which covers the most common and important disease-transmitting mosquitoes at the global scale, with a total of 9,900 mosquito images. The image resolution was 4464 × 2976 pixels. The specific taxonomic status and corresponding images are shown in Fig. 2. Due to the limitation of field collection, Ae. vexans, Coquillettidia ochracea, Mansonia uniformis, An. vagus and Toxorhynchites splendens only have females or only have males. In addition, each mosquito species included 300 images of both sexes, which was large enough and same number for each species, in order to balance the capacity and variety of training sets.Figure 2Taxonomic status and index of mosquito species included in this study Both male and female mosquitoes were photographed from different angles such as dorsal, left side, right side, ventral side, etc. Except for 5 species, each mosquito includes 300 images of both sexes, and the resolution of mosquito photos were 4464 × 2976. Cx. pipiens quinquefasciatus, Cx. pipiens pallens, and Cx. pipiens molestus (subspecies level, in dark gray background) were 3 subspecies in Cx. pipiens Complex (species level).Full size imageWorkflow for mosquito species identificationA three-stage flowchart of building best deep learning model for identification of mosquito species model was adopted (Fig. 3). The first learning stage was conducted by three CNNs (the Mask R-CNN, DenseNet, and YOLOv5) and three transformer models (the Detection Transformer, Vision Transformer, and Swin Transformer). Based on the performance of the first-stage model and the real mosquito labels, the second learning stage involved adjusting the model parameters of the three Swin Transformer variants (T, B, and L) to compare their performances. The third learning stage involved testing the effects of inputting differently sized images (384 × 384 and 224 × 224) to the Swin Transformer-L model; finally, we proposed a deep learning model for mosquito species identification (Swin MSI) to test the recognition effects of different mosquito species. The model was validated on different mosquito species, with a focus on the identification accuracy of three subspecies within the Cx. pipiens Complex and the detection effect of novel mosquito species.Figure 3Flowchart of testing deep learning model for intelligent identification of mosquito species.Full size imageComparison between the CNN model and Transformer model results (1st round of learning)Figure 4a shows the accuracies obtained for the six different computer vision network models tested on the mosquito picture test set. The test results show that the transformer network model had a higher mosquito species discrimination ability than the CNN.Figure 4Comparison of mosquito recognition effects of computer vision network models and variants. (a) Comparison of mosquito identification accuracy between 3 CNNs and 3 Transformer; (b) The best effect CNN (YOLOv5) training set loss curve(blue), validation set loss curve(green) and validation set accuracy curve(orange); (c) The best effect Transformer (Swin Transformer) training set loss curve, validation set loss curve and validation set accuracy curve. (d) Swin-MSI-T test result confusion matrix; (e) Swin-MSI -B test result confusion matrix; (f) Swin-MSI -L test result confusion matrix. Confusion matrix of mosquito labels in which odd numbers represent females and even numbers represent males. The small squares in the confusion matrix represent the recognition readiness rate, from red to green, the recognition readiness rate is getting higher and higher An. sinensis: 1, 2; Cx. pipiens quinquefasciatus: 3, 4; Cx. pipiens pallens: 5, 6; Cx. pipiens molestus: 7,8 Cx. modestus: 9,10; Ae. albopictus: 11, 12 Ae. aegypti: 13, 14; Cx. pallidothorax: 15, 16 Ae. galloisi: 17,18 Ae. vexans: 19, 20; Ae. koreicus: 21, 22 Armigeres subalbatus: 23, 24; Coquillettidia ochracea: 25, 26 Cx. gelidus: 27, 28 Cx. triraeniorhynchus: 29, 30 Mansonia uniformis: 31, 32 An. vagus: 33, 34 Ae. elsaie: 35,36 Toxorhynchites splendens: 37, 38.Full size imageIn the CNN training process (applied to YOLOv5), the validation accuracy requires more than 110 epochs to grow to 0.9, and the validation loss requires 110 epochs to drop to a flat interval; in contrast, during the training step, these losses represent a continuously decreasing process. These results indicate that the deep learning model derived based on the Swin Transformer algorithm was able to achieve a higher recognition accuracy in less time than the rapid convergence ability of the CNN during the iterative process (Fig. 4b).The Swin Transformer model exhibited the highest test accuracy of 96.3%. During the training process, the loss of this model could stabilize after 30 epochs, and its validation accuracy could grow to 0.9 after 20 epochs; during the validation step, the loss can drop to 0.36 after 20 epochs, after which the loss curve fluctuated but did not produce adverse effects (Fig. 4c). Based on the excellent performance of the Swin Transformer model, this model was used as the baseline to carry out the subsequent analyses.Swin Transformer model variant adjustment (2nd round of learning)Following testing performed to clarify the superior performance of the Swin Transformer algorithm, we chose different Drop_path_rate, Embed_dim and Depths parameter settings and labeled the parameter sets as the Swin Transformer-T, Swin Transformer-B, and Swin Transformer-L variants. Drop_path is an efficient regularization method, and an asymmetric Drop_path_rate is beneficial for supervised representation learning when using image classification tasks and Transformer architectures. The Embed_dim parameter represents the image dimensions obtained after the input red–green–blue (RGB) image is calculated by the Swin Transformer block in stage 1. The Depths parameter is the number of Swin Transformer blocks used in the four stages. The parameter information and test results are shown in Table 1. Due to the increase in the Swin Transformer block and Embed_dim parameters in stage 3, the recognition accuracies of the three variants were found to be 95.8%, 96.3%, and 98.2%, Correspondingly, the f1 score were 96.2%, 96.7% and 98.3%; thus, these variants could effectively improve the mosquito species identification ability in a manner similar to the CNN by increasing the number of convolutional channels to extract more features and improve the overall classification ability. In this study, the Swin Transformer-L variant, which exhibited the highest accuracy, was selected as the baseline for the next work.Table 1 Parameters and test accuracy of three variants of Swin Transformer.Full size tableBy plotting a confusion matrix of the test set results derived using the three Swin Transformer variants, we clearly obtained the proportion of correct and incorrect identifications in each category to visually reflect the mosquito species discrimination ability (Fig. 4d–f). In the matrix, the darker diagonal colors indicate higher identification rate accuracies of the corresponding mosquito categories. Among them, five mosquito species were missing because the Ae. vexans, Coquillettidia ochracea, Mansonia uniformis, An. vagus and Toxorhynchites splendens species were represented in the dataset by only females or only males. The confusion matrix shown in Panel C lists the lowest number of mosquito species identification error points and the lowest accuracy level obtained in each category, suggesting that the Swin Transformer-L model has a better classification performance than the Swin Transformer-T and Swin Transformer-B models.Effect of the input image size on the discrimination ability (3rd round of learning)To investigate the relationship between the input image size and mosquito species identification performance, in this study, we conducted a comparison test between input images with sizes of 224 × 224 and 384 × 384, based on the Swin Transformer-L model, and identified 8 categories of mosquito identification accuracy differences. These test results are shown in Table 2. When using an image size of 224 × 224 pixels, the batch_size parameter was set to 16, and when using an image size of 384 × 384 pixels, the batch_size parameter was set to 4; under these conditions, the proportion of utilized video memory accounted for 67%, as shown in Eq. 4, and this was consistent with the description of the relationship between the size of self-attentive operations during the operation of the Swin Transformer model when 384 × 384 pixels images were used. The time required for the Transformer-L model to complete all the training sessions was excessive, reaching 126 h and even exceeding the 124 h required by the YOLOv5 model, which was found to require the highest computation time during the training process in this work. Long-term training process could more fully reflect the performance differences between models. Fortunately and actually, the response speed of the model will not be affected by the training time. Compared to the accuracy of 98.2% obtained for 224 × 224 inputs, the 384 × 384 input image size derived based on the Swin Transformer-L model provided a higher mosquito species identification accuracy of 99.04%, representing an improvement of 0.84%.$$Omega ({text{W}} – {text{MSA}}) = 4{text{HWC}}^{2} + 2{text{M}}^{2} {text{HWC}}$$
(1)
Table 2 Comparison of recognition accuracy for different input image sizes.Full size tableVisualizing and understanding the Swin MSI modelsTo investigate the differences in the attentional features utilized by the Swin MSI and taxonomists for mosquito species identification, we applied the Grad-CAM method to visualize the Swin MSI attentional areas on mosquitoes at different stages. Because the Swin Transformer has different attentional ranges among its multi-head self-attention steps in different stages, different attentional weights can be found on different mosquito positions. In stage 1, the feature dimension of each patch was 4 × 4 × C, thus enabling the Swin Transformer’s multi-head self-attention mechanism to give more attention to the detailed parts of the mosquitoes, such as their legs, wings, antennae, and pronota. In stage 2, the feature dimension of each patch was 8 × 8 × 2C, enabling the Swin Transformer’s multi-head self-attention mechanism to focus on the bodies of the mosquitoes, such as their heads, thoraces, and abdomens. In stage 3, when the feature dimension of each patch was 16 × 16 × 4C, the Swin Transformer’s multi-head self-attention mechanism could focus on most regions of the mosquito, thus forming a global overall attention mechanism for each mosquito (Fig. 5). This attentional focus process is essentially the same as the process used by taxonomists when classifying mosquito morphology, changing from details to localities to the whole mosquito.Figure 5Attention visualization of representative mosquitoes of the genera Ae., Cx., An., Armigeres, Coquillettidia and Mansonia. This is a visualization for identifying the regions in the image that can explain the classification progress. Images of Toxorhynchites contain only males, with obvious differences in morphological characteristics, are not shown.Full size imageAe. aegypti is widely distributed in tropical and subtropical regions around the world and transmits Zika, dengue and yellow fever. A pair of long-stalked sickle-shaped white spots on both shoulder sides of the mesoscutum, with a pair of longitudinal stripes running through the whole mesotergum, is the most important morphological identification feature of this species. This feature was the deepest section in the attention visualization, indicating that the Swin MSI model also recognized it as the principal distinguishing feature. In addition, the abdominal tergum of A. aegypti is black and segments II-VII have lateral silvery white spots and basal white bands; the model also focused on these areas.Cx. triraeniorhynchus is the main vector of Japanese encephalitis; this mosquito has a small body size, a distinctive white ring on the proboscis (its most distinctive morphological feature), and a peppery color on its whole body. Similarly, the model constructed herein focused on both the head and abdominal regions of this species.An. sinensis is the top vector of malaria in China and has no more than three white spots on its anterior wing margin and a distinct white spot on its marginal V5.2 fringe; this feature was observed in Stage 2, at which time the modelstrongly focused on the corresponding area.The most obvious feature of Armigeres subalbatus is the lateral flattening and slightly downward curving of its proboscis; the observation of the attention visualization revealed that the constructed model focused on these regions from Stage 1 to Stage 3. The mesoscutum and abdominal tergum were not critical and were less important for identification than the proboscis, and the attention visualization results correspondingly show that the neural network focused less on these features.Coquillettidia ochracea belongs to the Coquillettidia genus and is golden yellow all over its body, with the most pronounced abdomen among the analyzed species. The model showed a consistent morphological taxonomic focus on the abdomen of this species.Mansonia uniformis is a vector of Malayan filariasis. The abdominal tergum of this species is dark brown, and its abdominal segments II-VII have yellow terminal bands and lateral white spots, which are more obvious than the dark brown feature on proboscis. Through the attention visualization, we determined that the Swin MSI model was more concerned with the abdominal region features than with the proboscis features.Subspecies-level identification tests of mosquitos in the Culex pipiens ComplexFine-grained image classification has been the focus of extensive research in the field of computer vision25,26. Based on the test set (containing 270 images) constructed herein for three subspecies of the Cx. pipiens Complex, the subspecies and sex identification accuracies were 100% when the Swin MSI model was used.The morphological characteristics of Cx. pipiens quinquefasciatus, Cx. pipiens pallens, and Cx. pipiens molestus within the Cx. pipiens Complex are almost indistinguishable, but their host preferences, self-fertility properties, breeding environments, and stagnation overwintering strategies are very different27. Among the existing features available for morphological classification, the stripes on the abdominal tergum of Cx. pipiens quinquefasciatus are usually inverted triangles and are not connected with the pleurosternums, while those of Cx. pipiens pallens are rectangular and are connected with the pleurosternums. Cx. pipiens molestus is morphologically more similar to Cx. pipiens pallens as an ecological subspecies of the Cx. pipiens Complex. However, taxonomists do not recommend using the unstable feature mentioned above as the main taxonomic feature for differentiation. By analyzing the attention visualization results of these three subspecies (last three rows on Fig. 5), we found that the neural networks of Cx. pipiens quinquefasciatus, Cx. pipiens pallens, and Cx. pipiens molestus still focused on the abdominal regions, as shown in dark red. The area of focus of these neural networks differ from that of the human eye, and the results of this study suggest that the Swin MSI model can detect finely granular features among these three mosquito subspecies that are indistinguishable to the naked human eye.Novel mosquito classification attributionAfter we performed a confidence check on the successfully identified mosquito images in the dataset, the lowest confidence value was found to be 85%. A higher confidence threshold mean stricter evaluation criteria, which can better reflect the powerful performance of the model. Therefore, 0.85 was set as the confidence threshold when judging novel mosquitoes. When identifying 10 unknown mosquito species, the highest derived species confidence level was below 85%; when the results were output to the genus level (Fig. 1d), the average probability of obtaining a correct judgment was 96.26%accuracy and 98.09% F1-score (Table 3). The images tested as novel Ae., Cx. and An. mosquito were from Minakshi and Couret et al.28,29.Table 3 Probability of correct attribution of novel species.Full size table More

Most records of N-tiger cats were from savanna environments, and it was not surprising that this vegetative formation has a key influence on the N-tiger cat range in the Amazon. The bulk of the L. t. tigrinus distribution lies in the savannas, dry forests and shrublands of the Cerrado and Caatinga biomes. These are also the areas with the vast majority of records for this lowland subspecies (Supplemental Fig. S5). Hence, L. t. tigrinus is more associated with savannas and savanna-like environments than with rainforests. In fact, more than 80% of the records in the Amazon were within 100 km of a savanna patch. Colonization of the northern savanna formations of the Amazon by the N-tiger cat likely occurred during the forest-savanna shifts of the glacial period18, and the cat currently shows a patchy distribution. Strong evidence of established biogeographic corridor connections between the savannas of the Cerrado and those of the Amazon exists, suggesting northward expansion of the former during glacial periods, perhaps predating the Last Glacial Maximum19,20,21. Further corroborating this evidence, tiger cat ‘gene flow’ niche modelling showed prior connectivity between the Guiana population and that of Central Brazil and no connectivity with the Andean population22. Additionally, Guianan tiger cat skin patterns are found in savanna and transitional savanna/Amazon areas and in the semiarid shrub-woodland of Brazil and are very distinct from the patterns of the tiger cats from the Andes of northwestern South America and Central America (Supplementary Information Fig. S6).The bioclimatic variables in the best model also supported the cat’s preference for savanna areas. The best model indicated a positive effect of precipitation in the driest month on the probability of the presence of the N-tiger cat, likely indicating the Aw/As climates of tropical savannas23. These climates are marked by seasonal variation in rainfall, with a pronounced dry season. Higher rainfall during the dry season favors the growth of vegetation, which results in some tree cover within the savannas. Thus, our results agree with previous research suggesting that tiger cats avoid open savanna formations24. Similarly, the species had a significant negative response to net primary productivity. This also supports the species’ avoidance of dense lowland rainforests, which are the most productive habitats. In the Amazon biome, the least productive areas are found in more open landscapes25.The N-tiger cat’s range considered from an ecoregion perspective12 could biogeographically explain its distribution in the Amazon. All records but 2 fell within Guiana savannas, Guiana highland forest, Guiana rainforest, part of the Uatumã-Trombetas rainforest bordering the Guianas or all of it connecting to Gurupá and Monte Alegre varzea forests, as well as Marajó varzeas, the interfluve Tocantins-Araguaia/Maranhão, and the southern block of the interfluve Xingu/Tocantins-Araguaia. There were two records from the Negro-Branco moist forest, which also includes savanna-like “campinarana” formations. The range also reaches the transitional babaçu palm forests of Maranhão and the Mato Grosso seasonal forests (Supplementary Information Fig. S7, Table S3). The N-tiger cat’s range in the Amazon was determined by combining records with species distribution modeling, also matching the ecoregion perspective.Outside the Guiana Shield and likely the savanna patches of the region of the Upper Negro River, in other parts of the Amazon, the N-tiger cat seems to be restricted to the forests of the eastern Amazon, along the arc of deforestation and to transitional areas with savanna formations. The presence and absence points at camera-trapping sites could explain the N-tiger cat’s range in the Amazon and define its distribution range in the biome. Absence points, for instance, were usually located in dense rainforest habitats throughout the Amazon biome.The species may occasionally occupy rainforests, such as those of the Guianas, where it tends to be very rare. At a site in central Suriname, after an enormous trapping effort of  > 20,000 trap days in four years by cat specialists, over an area  > 1100 km2, no records of the N-tiger cat were found (Supplementary Information Table S2), although its presence is expected in that area26. This finding attests to the inherent rarity of this felid in its limited range within the Amazon. However, could its association with the arc of deforestation be related to the replacement of forest by bushy savanna-like vegetation that succeeds abandoned pastures? The other currently recognized subspecies, L. t. pardinoides (the Andean tiger cat) and L. t. oncilla (the oncilla), and the recently split southern tiger cat L. guttulus are all associated with forested areas. Conversely, L. t. tigrinus has higher abundance and is mostly found in the nonforested habitats of the Cerrado and Caatinga domains of Brazil and only rarely in rainforests. Thus, L. t. tigrinus may be an open-habitat (sub)species. However, within savannas, N-tiger cats are restricted to denser savanna formations, with open savannas deemed unsuitable24. In the semiarid Caatinga, the N-tiger cat also prefers denser formations27,28.One of the most interesting findings was the clear relationship between the ranges of the dominant mesopredator and subordinate species. The ranges of ocelots and N-tiger cats in the Amazon were diametrically opposite (Fig. 1), a finding never recorded for felids. The reported ocelot densities and relative abundance indexes (RAIs) in the Amazon range from 0.29 to 0.95 ind/km2 and 0.07–13.2 ind/100 trap-days, respectively7,29. Thus, the expected ocelot density found using modeling that allows for N-tiger cat presence is very low (Fig. 2A). In the Rupununi, the ocelot:N-tiger cat RAI ratio was roughly 10:1, with a very low RAI and expected density for N-tiger cats (see Supplementary Material). The only other relative abundance estimate of tiger cats presented for the Amazon30 was not confirmed as an estimate of tiger cats following inspection of the original records by the authors but as an estimate of margays or ocelots. This antagonistic relationship between ocelots and all other small cat species in their area of sympatry is quite impressive. It is density-dependent, as it seems to take effect only above an ocelot density threshold of 0.12 ind./km231. The influence can range from patterns of density, distribution, and occupancy to spatial and temporal use. Conversely, such an impact was not detected when either the small cats or ocelots were compared to the larger cats31,32,33,34,35.In view of the Red List assessments and applying the limited estimates presented, the expected total population size for N-tiger cats in the Amazon would be approximately 150 and 1622 individuals, considering their AOO or EOO, respectively. Applying the IUCN’s formula for mature individuals8, these numbers would be 45 and 487 individuals for the AOO and EOO, respectively.The ocelot’s preference for very dense rainforests may explain the low probability of N-tiger cat occurrence within the Amazon biome. Notably, most tiger cat records from rainforests and all those from premontane forests came from the Guiana Shield, a region where tropical grasslands and savannas dot more forested landscapes. The Guiana Highlands and Pantepui ecoregions, which make up a considerable portion of the shield, tend to have low ocelot densities (below 0.30 ind/km2), although they do contain some rainforest. Ocelot densities reach some of their lowest values in the Guianan savanna ecoregion (mean ocelot density of 0.029 in the savanna formations), where the N-tiger cat probability of occurrence was highest. At the Karanambu site in the Rupununi, all ocelot records came from either gallery forests or forest patches embedded in the savanna. Although the data did not allow us to test further hypotheses, it is likely that spatial partitioning occurs in the Guiana Shield, with N-tiger cats favoring habitats that are more open. Conversely, areas farther west in the Amazon biome, other than the predicted area, do not have any major savanna patches and are covered mostly by lowland tropical rainforest formations, where ocelots can potentially reach densities in excess of 0.7 ind/km2. Of all Amazonian records of N-tiger cats, only one came from west of the 68th meridian: a preserved specimen from Puerto Leguizamo on the Putumayo River in Colombia. The specimen was identified as L. t. pardinoides by its collector, so it most likely represents an individual that came down from the foothills of the Andes. Alternatively, it could have been caught in the Andean foothills but labeled generally as from Puerto Leguizamo, as museum records do not always present precise locations, like most of those from our dataset; thus, they could represent a broader region, not a single collection location.The records of L. t. tigrinus in the Monte-Alegre Várzea ecoregion and Tapajós-Xingu Moist Forest ecoregion (which shares a border with the Amazon River) are actually from the small savanna patches of Terra Santa and Alter do Chão, respectively, which are imbedded within the forests of these ecoregions. Similarly, the Negro-Branco Moist Forest ecoregion includes open-canopy white sand forests with savanna-like vegetation, known as ‘campinaranas’36.Although our model predicted a high probability of N-tiger cat presence in the Marajó Várzea ecoregion, the records from the island came from savanna patches and not from flooded forests and mangroves. Hence, we did not include such large areas in the AOO for the subspecies. It is likely that the highly predicted probability of presence there is an artifact of low predicted ocelot density. Nevertheless, the environment there is not suitable for either cat. Our ocelot density model was highly significant and explained almost 50% of the variation in ocelot density. The remaining variation was related to either other variables that could not be measured via satellite imagery (such as prey availability) or the sampling design of the different studies. Nonetheless, ocelot densities predicted from our model across the Amazon were within the expected range for the species29.Why are N-tiger cats absent in camera-trapping studies in Amazonian forests throughout the biome? The most straightforward answer seems to be because they simply are not there (central and western Amazon) or, where present, their numbers are extremely low (Guianas and eastern Amazon). The lack of surveys cannot be cited as a potential reason for their apparent absence because the studies that did not detect the species were conducted throughout the Amazon biome, in all nine Amazonian countries. Some of the areas have been surveyed for several years—or decades in some cases—and have failed to record a single individual (Supplementary Information Table S2). Typically, N-tiger cats appear, even prominently, on cameras in other biomes, such as in the savannas of the Cerrado and semiarid scrub of the Caatinga domain in Brazil, including sites where ocelots are present24,27,37. Clouded tiger cats (L. t. pardinoides) have also been frequently recorded on cameras in the Andes, higher than 1500 m above sea level34,38, but not in lowland Amazonian forests. This finding indicates that the N-tiger cat is not camera-shy. In northern Brazilian savannas, its density can reach 0.25 ind/km2 24. Coincidentally, this highest density estimate of the N-tiger cat is the same as the lowest ocelot density estimate for Amazonian forests24,29.Tiger cats and margays show high similarity, making misidentifications relatively common39. However, the evaluation of  > 3000 camera trap images of small-medium felids in the Amazon revealed that only one mildly resembled a tiger cat, a finding that supports the species being absent there and does not represent a case of mistaken identity with margays or even ocelots7.The Amazonian range of L. tigrinus is very limited, and populations are expected to be very small. With the upcoming split of L. t. tigrinus and L. t. pardinoides into two different species40, this situation would have serious implications for the conservation of the former. Thus, L. t. tigrinus conservation lies outside the “Amazonian safe haven” of most other carnivore species found there7. The Brazilian drylands Cerrado and Caatinga represent such places for L. t. tigrinus populations. Unfortunately, these biomes have had  > 50% of their cover completely removed41. Very importantly, besides being extremely rare in the Amazonian savannas, this rather limited vegetative formation is also considered highly threatened and of conservation priority42. Therefore, the tiger cat could become an emblematic flagship species representing the uniqueness of this vegetative formation in dire need of protection.In short, the picture that emerges is that although the N-tiger cat uses both rainforests and deciduous forests in the Amazon, it seems to be mostly associated with savanna formations and that its distribution in the Amazon is highly influenced by the ocelot, the dominant mesopredator. The N-tiger cat’s inherent rarity, expected population size, and restricted range in the Amazon suggest that this biome does not in fact represent a safe haven for the global conservation of this small felid. In addition to shedding light on and refining the N-tiger cat distribution in the Amazon, this paper highlights the importance of including biological variables, such as the potential impacts of competitors and predators on species presence and distribution, in SDMs. More

The richness of Melanesian FrogsApproximately 7.2% (534 out of 7404) of Earth’s recognised frog species occur in Melanesia, a region comprising < 0.7% of the world’s land area. Frog richness in Melanesia, and especially on New Guinea and nearby land-bridge islands (471 species), is higher than in any other tropical insular region (Fig. 1a). New Melanesian frog species have been described at an average rate of nearly 13 species/year since 2000, and the recognised frog fauna has grown by > 50% in that timeframe (Fig. 1b). The authorship of new species has been concentrated, with six authors featuring on 20 or more descriptions since 2000, and one or more of these six authors on every species description since 2000. A small number of species descriptions has included genetic data (31 species), although a higher number of Melanesian frog species have at least one sequence available on GenBank (~38%, or approximately 200 species). This taxonomic work has revealed or emphasised many evolutionary novelties (Fig. 2): multiple apparently independent derivations of extremely miniaturised vertebrates22,23,24, including some of the world’s smallest known tetrapods23,25,26; multiple derivations of complex parental care in different genera27,28; frequent evolutionary shifts between terrestrial, arboreal and scansorial lifecycles22,29; the most extreme sexual size dimorphism yet documented in anurans30; drastic ontogenetic colour change31; a radiation of canopy-dwelling treefrogs32 that show extensive finger webbing and parachuting behaviour convergent with unrelated frog lineages in Asia and the Neotropics; and treefrogs with erectile noses33,34. Taxonomic work has also elucidated novel concentrations of range-restricted endemic taxa, especially in the Milne Bay Region at the far eastern edge of New Guinea21.Fig. 1: Temporal trends in the documentation of the Melanesian frog fauna.a Species accumulation curves for species-rich ( >100 species) insular frog biotas (Species lists from AmphibiaWeb as of 1 October 2021). Separate accumulation curves are given for the entire fauna of Melanesia (including New Guinea), and the fauna of New Guinea and nearby predominantly land-bridge islands. b Species accumulation curve for frogs within Melanesia. Bar at end indicates predicted number of species in each major family based on known, but as yet undescribed candidate species.Full size imageFig. 2: Melanesian frog species described within the last 15 years illustrating the ecological and morphological diversity of the fauna.a Paedophryne titan and b Choerophryne gracilirostris – examples of lineages that have undergone convergent minaturisation; c Choerophryne alpestris – a fossorial species within a largely scansorial lineage; d Xenorhina macrodisca – scansorial species within a largely fossorial lineage; e Cornufer custos and f Oreophryne oviprotector – independent derivations of complex parental care; g Litoria pallidofemora – extensive digital webbing for parachuting; and h Litoria pinocchio – sexually dimorphic and erectile rostral spikes. Photographs F. Kraus (a), S. Richards (b–g), and courtesy of T. Laman (h).Full size imageFrog species richness in Melanesia is highly concentrated into just three families, with Pelodryadidae (137 recognised species, estimated ~200) and especially Microhylidae (317 species, estimated >400) dominating. Melanesian Pelodryadidae are phylogenetically interdigitated with relatives in Australia, suggesting multiple dispersals between the two regions35. In contrast, ancestors of the direct-developing microhylids colonised Melanesia from Asia via trans-marine dispersal likely only once36, radiated across open ecological niches37, and are now the most species-rich insular radiation of frogs in the world. The third major family comprises an ecologically diverse radiation of the direct-developing Ceratobatrachidae (57 species, estimated 66) largely associated with island-arc terranes of East Melanesia and the Philippines, indicating a long history of insular diversification and trans-marine dispersal38. The predominance of direct-developing frogs in Melanesia (~70% of species) mirrors insular faunas in Madagascar (~34%), Sri Lanka (~67%) and the Greater Antilles (~87%). The other four frog families in Melanesia are all relatively species poor (2, 3, 4, and 13 species) (Fig. 1a), centred in New Guinea, and include lineages originating in Asia (Ranidae, Dicroglossidae) or Australia (Myobatrachidae, Limnodynastidae).The described diversity of Melanesian amphibian species remains an underestimate. Survey work and investigation of museum collections by the co-authors identified ~190 additional candidate species distributed across 16 different genera, mostly from Papua New Guinea, suggesting a total richness of over 700 frog species (Fig. 1a, Supplementary Table 1). This estimated percentage of undescribed diversity (~25%) mirrors estimates for the New Guinean flora (~18–22%)7. The majority of candidate species are concentrated in the two most diverse families (Microhylidae and Pelodryadidae), although genetic, morphological, and acoustic evidence indicate the diversity of Melanesian Ranidae is also underestimated (S. Richards and F. Kraus pers. obs.). Most material documenting candidate species has been collected in the last 20 years, and the vast majority is from Papua New Guinea (Supplementary Fig. 1). There is some suggestion of a slowing in the rate of candidate species discovery in the last decade (Supplementary Fig. 2); however, several of the most active field workers in this region have ceased survey work in recent years, which likely accounts for much of this decline. The pervasiveness of complexes of morphologically and/or acoustically cryptic taxa is poorly understood; survey work continues to reveal novelties, and large areas of the region remain unsurveyed or undersampled. In particular, comparisons of area-to-diversity ratios between the better-known eastern portion of New Guinea (Papua New Guinea) with the poorly surveyed western (Indonesian) portion of the island further suggest that, even with candidate species included, diversity in the latter region may be underestimated by as much as 50% (Supplementary Methods and Results, Supplementary Table 2). These trends and patterns all indicate that ~ 700 species is a very conservative minimum estimate of total diversity and support analyses in other taxa showing Melanesia remains a hotspot of unrecognised diversity39,40.Endemism and distributional patternsThe Melanesian frog fauna is highly endemic (97.2%), with tiny proportions of species shared with Australia (2.4%) or with islands farther west in Indonesia (0.6%), indicating that Australia and Melanesia are discrete centres of frog diversification, despite periodic connection via land bridges through the late Tertiary41. The vast majority of Melanesian frog species (471) occur on New Guinea and nearby land-bridge islands (Raja Ampats, Japen and the Milne Bay islands). In comparison, the frog fauna of the much smaller region of Maluku is depauperate (16 species, of which nine are endemic) but also almost certainly underestimated (e.g., there are no Microhylidae recorded from Buru). Most taxa from Maluku are congeneric (and several conspecific) with lineages centred on New Guinea, supporting the biogeographic clustering of Maluku’s amphibians with the main island of New Guinea. In contrast, the frog fauna of East Melanesia is more diverse and highly endemic and dominated by an ecologically diverse radiation of a different family (Ceratobatrachidae) with only four (all pelodryadid treefrogs) out of 56 species shared with nearby New Guinea. East Melanesia and New Guinea appear to be discrete and long-isolated centres of diversification, as expected from their independent geological histories42.Melanesia spans five countries, and this has possibly to some degree masked the exceptional species diversity of the overall region. Papua New Guinea has the highest number of species (398) and endemic species (318). This likely reflects some combination of its slightly larger area (when islands to the north are included), more diverse geological origins, and greater inventory work than seen in neighbouring regions of Indonesia7. Papua, West Papua and Maluku (Indonesia) have many fewer documented species (197), of which a majority (134) is endemic. The boundary between Papua and Papua New Guinea is visible in species-richness maps (Fig. 3a), with lower diversity to the west, indicating that the distribution and diversity of frogs in Indonesia remain less documented in science. The frog faunas of the Solomon Islands (21 species) and Fiji (two species) are more depauperate but include a significant endemic or near-endemic component, whereas the geographically intervening islands of Vanuatu support no native frogs.Fig. 3: Frog species richness in Melanesia based on IUCN distributional maps for all species described by 2019.a All species; b Ceratobatrachidae; c Microhylidae; d Pelodryadidae. Areas of highest estimated diversity correspond to mountain ranges in central and northern New Guinea. The boundaries between Maluku, New Guinea and East Melanesia are indicated. Fiji has only two frog species and is geographically distant from other areas of Melanesia inhabited by frogs and is not visble on this map.Full size imageBased on distribution maps generated for all species recognised by 31 August 2019, the highest regional alpha diversity of frogs occurs along the Central Cordillera of New Guinea (especially in Papua New Guinea) and around the higher mountain ranges along the north coast of Papua New Guinea (Fig. 3a). These centres of diversity correlate with extensive areas of hill and montane forest and broadly correspond with elevational species-richness patterns for mammals and birds in Melanesia43 and for many other taxa elsewhere in the tropics44,45. Large areas of montane forest with lower species richness along the northern versant of the Central Cordillera in Papua New Guinea and in mountain ranges across Papua certainly reflect inadequate sampling. The ceratobatrachid-dominated frog fauna of East Melanesia is richest in Bougainville (Fig. 3d), with attenuating richness towards the west and especially to the east. The two most speciose families both show alpha diversity peaks in mountainous areas of central New Guinea (Fig. 3c–d). In contrast, microhylids are largely absent from the seasonally dry woodlands of the Trans-Fly region in southern New Guinea and exhibit high diversity in northern New Guinea, whereas pelodryadids are much more speciose in the lowlands of southern New Guinea than northern New Guinea. These broad trends may have both ecological (sensitivity of direct-developing microhylids to dry conditions) and historical (Australia as a centre of origin for savanna-adapted Pelodrydidae) underpinnings.The historical and contemporary factors underpinning high frog species diversity in New Guinea remain largely unstudied, especially when compared to other species-rich insular amphibian faunas such as Madagascar46 or the Greater Antilles47. When compared to some areas of the Neotropics, alpha and beta diversities of frogs in lowland forests in the basins of the Sepik and Ramu rivers in New Guinea are unremarkable48. However, the Milne Bay Region has exceptionally high levels of endemism21, so species turnover will be higher in this area. Extent-of-occurrence estimates derived from IUCN maps indicate that direct-developing microhylids have smaller mean and median range sizes than all other families of frogs in Melanesia (Supplementary Table 3). Microhylidae also dominate anuran species diversity in Milne Bay21 and many mountain areas where standing water is very limited49. These data suggest that, as with some areas in the Neotropics50, high beta diversity in lineages with direct development is a key factor underpinning amphibian megadiversity in Melanesia. To address these questions further, synthetic analyses are required to better quantify the extent to which regional megadiversity in Melanesia reflects high community diversity versus species turnover, how elevation and insularity moderates these two parameters, and to what extent emergent patterns may differ from diverse frog communities in other regions such as the Neotropics.The conservation status of Melanesian FrogsThe frog fauna of Melanesia is currently less threatened but more Data Deficient than other comparable insular regions (Fig. 4a). The vast majority of Melanesian frogs are categorised as Least Concern (68%) or Data Deficient (24%). Thirty-one species (6%, or 8% if Data Deficient taxa are excluded) are threatened (Critically Endangered, Endangered, Vulnerable) (Supplementary Table 4), and eight species are considered Near Threatened. No species are assessed as Extinct or Extinct in the Wild. Since the first Global Amphibian Assessment in 2004, the number of Melanesian frog species has grown by 44%, and nearly 60% of the 31 Melanesian frog species now considered threatened were described after 2004 (Fig. 1a). Only one change in status between 2004 and 2019 was considered genuine (Cophixalus sphagnicola), due to the emerging threat of a newly opened mine. All other status changes (for 116 taxa) reflect better information on distribution or changed assessment protocols (Supplementary Table 5). Applying stricter criteria for use of the Data Deficient category in the 2019 IUCN assessment reduced the number of Data Deficient species when compared to 2004 (125 versus 197), but Melanesia still has a higher percentage of Data Deficient taxa than other species-rich tropical insular faunas (Fig. 4a).Fig. 4: The conservation status of Melanesian frogs.a Comparison of number of species in each IUCN threat category across Melanesia, other diverse insular regions, and the nearby continent of Australia. Melanesia has a proportionally low number of threatened taxa but high number of Data Deficient taxa (EX Extinct, CR Critically Endangered, EN Endangered, VU Vulnerable, NT Near Threatened, DD Data Deficient, LC Least Concern, NE Not Evaluated); b Slopes around Mt Simpson, Milne Bay Province, a hotspot of threatened frog diversity due to forest loss through conversion to anthropogenic grasslands; c Choerophryne sanguinopicta from Mt Simpson (Critically Endangered); d Oreophryne ezra from Rossel Island (Critically Endangered) and; e Cornufer citrinospilus from New Britain (Vulnerable). Photographs F. Kraus (b–d), S. Richards (e).Full size imageAll Critically Endangered and Endangered—and most of the Vulnerable—species were listed because of their small extent of occurrence and on-going decline in habitat area and/or quality (criteria B1ab(iii)) (Supplementary Table 6). The key threatening processes were typically forest disturbance or loss due to conversion to plantations or gardens, repeated burning, or mining (Fig. 4b–c). Only two insular species with very localised montane distributions were considered threatened by climatic disturbance and/or climate change alone (Cornufer citrinospilus and Oreophryne ezra) (Fig. 4d–e). No species were currently declining from pathogens, and in particular Batrachochytrium dendrobatidis (Bd), which remains undetected in Melaneisa51. However, the introduction and establishment of Bd has been identified as a severe threat for well over one hundred taxa52, especially for montane pelodryadid treefrogs, a group that has been devastated by this disease in parts of Australia.Although much of New Guinea has historically been considered a ‘wilderness area’ with comparatively little human impact53, the distributions of threatened taxa also highlight areas of conservation concern wherein range-restricted (often single-island endemic) taxa overlap with extensive and increasing anthropogenic impacts (Fig. 5a–b). Nearly half the species identified as threatened (13) are restricted to a recently delineated dramatic centre of herpetofaunal endemism in the Milne Bay Region at the eastern tip of Papua New Guinea21. Three clusters of small-range endemics in this region (all documented in the last two decades) present immediate conservation issues. The first is Mount Simpson, where six microhylids (four named, two awaiting description) with highly restricted ranges are threatened by habitat loss, especially repeated burning and associated conversion of forest to grassland (Fig. 4b). The second is Woodlark Island, where the status of seven endemic microhylids (six named, one undescribed) is likely to worsen rapidly if current, approved proposals to convert large areas of primary forest to oil-palm plantation and/or gold mines proceed21. Finally, Misima Island is home to four endemic microhylids (two considered threatened) with ranges that overlap areas disturbed by mining and forest loss21. Other regions with multiple overlapping threatened taxa are the Adelbert Mountains in Morobe Province (two species), New Britain (two lowland species and one highland species), and Greater Bukida in the Solomon Islands (three lowland species). These clusters of narrow-range taxa highlight important—and in most cases largely overlooked—conservation priorities for Melanesian frogs (and likely other taxa as well21,54). The high percentage of Data Deficient species and low level of survey effort in many areas (especially Papua and West Papua Provinces, Indonesia) also raise the possibility that other threatened hotspots remain overlooked. One area of particular concern may be the island of Biak in Indonesia, which has lost much of its primary vegetation but is home to at least three endemic frogs (one Data Deficient, two Least Concern).Fig. 5: The distribution of threatened frogs in Melanesia.a The estimated distribution of all 31 Melanesian frog species considered Critically Endangered, Endangered or Vulnerable at the end of 2019. Distributional areas are not colour coded by the number of threatened taxa. b Close up of the Milne Bay endemism hotspot. Distributional areas are colour coded by number of taxa, with darker tones indicating more taxa. In both a and b upland areas or islands where the distributions of two or more threatened species overlap are labelled and the number of threatened taxa are indicated in parentheses. Background maps uses the Shuttle Radar Topography Mission (SRTM) 30-meter digital elevation model, accessed from USGS Earth Explorer (https://earthexplorer.usgs.gov/).Full size imageUnderstanding and conserving a megadiverse biotaThe Melanesian flora and frog fauna are both now shown to be megadiverse and highly endemic, yet both also remain poorly known with large areas under-surveyed. An updated comprehensive assessment of threats and taxonomic trends across the frog fauna presented here further highlights that the biota of Melanesia remains relatively intact and less threatened when compared to other biodiverse insular regions. However, a large proportion of the fauna remains Data Deficient or undescribed, and key hotspots of endemism have been overlooked and are increasingly threatened. In both plants and anurans much scientific knowledge of Melanesia’s biota has also been contributed by a relatively small number of productive, but later-career researchers based outside of Melanesia7.Further documenting and conserving the exceptional diversity of Melanesia presents a suite of challenges and opportunities. Recommendations to enable improved documentation of plant megadiversity in Melanesia7 centre around training, capacity-building and support for taxonomy in Melanesia and globally, improving access to specimen collections and diagnostic resources, and ongoing support for survey and collecting within Melanesia. These recommendations apply equally to amphibians. However, addressing these challenges is tempered by the limited career opportunities available to ecologists and taxonomists (both in developed, but especially in developing countries), the variable quality of scientific infrastructure that exists across the region, and the high cost of doing fieldwork in remote areas with limited logistical infrastructure. In the context of these challenges, we hereby focus on suggesting some short-term key priorities and opportunities to build capacity for understanding and conserving frog biodiversity in Melanesia.First, over the last twenty years opportunities to employ Melanesian nationals in survey, monitoring and outreach work have been (and will continue to be) generated predominantly by NGOs, universities and large-scale extractive projects, for example through recent work in the gas fields of the Papua New Guinea Highlands49. While there are diverse perspectives on extractive industries, monitoring and survey work associated with large development projects are a key source of funds to provide training to enable Melanesians to undertake biodiversity work within the region. A key driver of this is strong environmental legislation required by some governments and major lending agencies, in particular the International Finance Corporation under Performance Standard 655. These requirements need to be maintained, enforced and, where possible, exceeded.To further support fieldwork by national scientists there is a need for more readily accessible identification resources for Melanesian researchers, land-owners and managers. An up-to-date comprehensive identification guide to the frog fauna of the whole region would assist and promote taxonomic, ecological and conservation research. However, for many Melanesians, small, regionally focused guides are more usable. These have already been produced for several areas (Supplementary References), providing a model that can be updated and transferred to other regions. Mobile phones are widely used throughout Melanesia, so app- and online-based identification resources may become increasingly accessible. Smartphone-friendly citizen science platforms like iNaturalist56 or even Facebook groups57 also provide potentially powerful resources through which locally collected data can be captured, vetted and disseminated, although their use is currently limited in Melanesia due to patchy internet coverage in many areas. Working with and supporting people from Melanesia to explore and increase the use of these resources could help to ensure longer-term preservation and accessibility of species records and associated data.The latest IUCN assessment for Melanesian frogs also highlights how taxonomic and conservation knowledge is accumulating rapidly. The key geographic areas of threat identified in our study were largely invisible to assessments made less than two decades ago (in 2004) both because the relevant taxonomic work had not been done, and because the situation in Melanesia is changing rapidly. To keep track of these rapid changes it is critical for workers in the region to work together to synthesise and collate new taxonomic, distributional and conservation data. Indeed, since the 2019 IUCN assessment over 20 additional species of Melanesian frogs have been described, and their conservation status should be assessed as a matter of urgency. Preliminary conservation assessments against IUCN criteria are increasingly being included in descriptions, and this trend should be supported and encouraged. More Melanesian nationals need to be involved in conservation assessment processes. Updated comprehensive conservation assessments of other vertebrate groups will also identify complementarity of conservation priorities among taxa in the Melanesian region.Patterns of distribution and threat suggest some geographic priority areas for documenting the diversity of amphibians (and potentially other low-vagility taxa) in Melanesia. First, work in eastern New Guinea has allowed the delineation of geographically localised clusters of threatened taxa that have until now gone unnoticed, perhaps in part because of the designation of much of Melanesia as a sparsely populated and comparatively undisturbed ‘wilderness’ area21. Most threatened frog taxa in these regions are associated with small islands or isolated ‘sky island’ mountains. The degree to which other taxa show endemism in these areas is poorly known. The biotas of potentially comparable islands in Indonesia such as the Raja Ampat Islands, Geelvink Bay and southern Maluku, also remain poorly known, suggesting additional priority areas for survey, taxonomic investigation and conservation assessment. Second, mid-elevation areas show highest alpha diversity, but large areas of this habitat, especially along the northern slopes of the Central Cordillera, remain poorly surveyed. The frog pathogen Bd has devastated montane communities of two Australian frog families that also occur in montane New Guinea (Myobatrachidae and Pelodryadidae)52. In the unfortunate event that Bd colonised New Guinea a wave of rapid declines and extinctions would likely follow52, so a strong baseline of information on montane species diversity, distributions and population status is critical for detecting these impacts. More

Sites overview and characteristicsThis study focused on three regions located in subantarctic, arctic, and subarctic latitudes. The respective latitudinal and longitudinal ranges covered in this study were: 54.95 to 52.08 °S, and 72.03 to 67.34 °W in Patagonia; 67.44 to 67.54 °N, and 86.59 to 86.71 °E in Siberia; 63.21 to 68.63 °N, and −150.79 to −145.98 °W in Alaska (Figs. 1 and 2). The exact coordinates for each sample were included in the submitted dataset. The field campaigns were conducted in 2016, during the summer for each respective region: January-February in Chilean Patagonia, June-July in Alaska and July-August in Siberia.Fig. 1Location of the three areas included in this study (panel a). The permafrost state and the number of sites and samples per region is indicated for each area. General views of 5 sites are provided as examples (b–f). Panel B provides a large view of the ecosystem surrounding the wetland ALP2 (Alaska, exact location indicated by the white circle). Lake PCL1 (panel c) is representative of the lakes on Navarino island (Chilean Patagonia). The glacial lake SIL2 is shown in panel d. At site SIP5, the hollow at first plan is surrounded by palsa (hummock, second plan), characterized by dark organic matter and lichen vegetation (panel e). The PPP3 peatland shown in panel f is dominated by Sphagnum magellanicum, like most peatlands in the area.Full size imageFig. 2Maps of sampling sites in Patagonia, Alaska and Siberia, indicating the ecosystem type (lake, wetland, soil). The tables show the complete- (in white) and the partial- (in grey) characterization sites. The exact coordinates of each sample are provided in the data record (See data records section).Full size imageFor every site included in the present study, a set of nine qualitative environmental and/or ecological site-scale descriptors was selected and adapted from ENVO Environment Ontology40, which included for example permafrost state, biome, environmental feature and vegetation type (Table 1, Fig. 3). Permafrost state was obtained from the NSIDC permafrost map41. The biome, large-scale descriptor based on climate and vegetation criteria, was derived from Olson et al.42. Temperate forest, boreal forest, and tundra biomes were included. The environmental features that were representative for the three regions were considered: lakes, wetlands, broadleaf/coniferous/mixed forest soils, grassland, tundra, and palsa. All the metadata was included in the submitted dataset. Table 2 summarizes the main types of sampled ecosystems and their main characteristics in the three regions, while Supplementary Table S1 provides the details of each sampling site.Table 1 Overview of the dataset contained in Mimarks sheet.Full size tableFig. 3Description of the qualitative environmental/ecological descriptors used to describe every sample, derived from ENVO Environment Ontology40.Full size imageTable 2 Main types of sampled ecosystems in the three studied regions.Full size tableIn Alaska, the studied area ranged from the Alaska Range and Fairbanks area (interior, continental climate, 63–65°N, discontinuous permafrost) up to Toolik Field Station (North Slope, arctic climate, 66–69°N, continuous permafrost; Fig. 2). The physiochemistry and CH4 emissions of lakes ALL1 (Killarney lake), ALL2 (Otto lake), ALL3 (Nutella lake), and ALL4 (Goldstream lake) were previously characterized35. A number of heterogeneous soil and wetland samples were collected around the studied Alaskan lakes and/or from monitored sites, as detailed in Supplementary Table S1. In the Alaska Range and Fairbanks area, soils were mostly covered by mixed or taiga forests, alpine tundra, and bogs or fens wetlands. In the norther Brooks Ranges mountain system, the landscape was piedmont hills with a predominant soil of porous organic peat underlain by silt and glacial till, all in a permafrost state, characterized mainly by Sphagnum and Eriophorum vegetation, as well as dwarf shrubs.In Siberia, the studied area was located in the discontinuous permafrost region surrounding Igarka, on the eastern bank of the Yenisei River (Fig. 2). This region was mainly covered by forest, dominated by larch (Larix Siberica), birch (Betula Pendula), and Siberian pine (Pinus Siberica), and palsa landscapes (frozen peat mounts), the latter being dominated by moss, lichens, Labrador tea and dwarf birch. In degraded areas, thermokarst bogs were dominated by Sphagnum spp. and Eriophorum spp. Land cover was an indicator of permafrost status, since forested areas reflected a deep permafrost table ( >2 m) associated with Pleistocene permafrost, while palsa-dominated landscapes were indicative of the presence of near-surface ( More

Hirai, H. et al. Organic micropollutants in marine plastics debris from the open ocean and remote and urban beaches. Mar. Pollut. Bull. 62(8), 1683–1692. https://doi.org/10.1016/j.marpolbul.2011.06.004 (2011).Article 
CAS 
PubMed 

Google Scholar 
Masó, M., Garcés, E., Pagès, F. & Camp, J. Drifting plastic debris as a potential vector for dispersing harmful algal bloom (HAB) species. Sci. Mar. 67(1), 107–111. https://doi.org/10.3989/scimar.2003.67n1107 (2003).Article 

Google Scholar 
Pandey, D., Singh, A., Ramanathan, A. & Kumar, M. The combined exposure of microplastics and toxic contaminants in the floodplains of North India: A review. J. Environ. Manag. 279, 111557. https://doi.org/10.1016/j.jenvman.2020.111557 (2021).Article 
CAS 

Google Scholar 
Peng, L. et al. Micro- and nano-plastics in marine environment: Source, distribution and threats—A review. Sci. Total Environ. 698, 134254. https://doi.org/10.1016/j.scitotenv.2019.134254 (2020).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Rillig, M. C. & Lehmann, A. Microplastic in terrestrial ecosystems and the soil?. Environ. Sci. Technol. 46(12), 6453–6454. https://doi.org/10.1021/es302011r (2012).Article 
CAS 
PubMed 

Google Scholar 
Rochman, C. M. & Hoellein, T. The global odyssey of plastic pollution. Science 368(6496), 1184–1185. https://doi.org/10.1126/science.abc4428 (2020).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Jan Kole, P., Löhr, A. J., van Belleghem, F. G. A. J. & Ragas, A. M. J. Wear and tear of tyres: A stealthy source of microplastics in the environment. Int. J. Environ. Res. Public Health https://doi.org/10.3390/ijerph14101265 (2017).Article 

Google Scholar 
Sommer, F. et al. Tire abrasion as a major source of microplastics in the environment. Aerosol Air Qual. Res. 18(8), 2014–2028. https://doi.org/10.4209/aaqr.2018.03.0099 (2018).Article 
CAS 

Google Scholar 
Beita-Sandí, W., Selbes, M., Ersan, M. S. & Karanfil, T. Release of nitrosamines and nitrosamine precursors from scrap tires. Environ. Sci. Technol. Lett. 6(4), 251–256. https://doi.org/10.1021/acs.estlett.9b00172 (2019).Article 
CAS 

Google Scholar 
Kaminsky, W. & Mennerich, C. Pyrolysis of synthetic tire rubber in a fluidised-bed reactor to yield 1,3-butadiene, styrene and carbon black. J. Anal. Appl. Pyrolysis 58–59, 803–811. https://doi.org/10.1016/S0165-2370(00)00129-7 (2001).Article 

Google Scholar 
Sundt, P., Schulze, P. E. & Syversen, F. Sources of microplastic- pollution to the marine environment. Mepex Nor. Environ. Agency 86, 20 (2014).
Google Scholar 
White, W. C. Butadiene production process overview. Chem. Biol. Interact. 166(1–3), 10–14. https://doi.org/10.1016/j.cbi.2007.01.009 (2007).Article 
CAS 
PubMed 

Google Scholar 
Alimi, O. S., Farner Budarz, J., Hernandez, L. M. & Tufenkji, N. Microplastics and nanoplastics in aquatic environments: Aggregation, deposition, and enhanced contaminant transport. Environ. Sci. Technol. 52(4), 1704–1724. https://doi.org/10.1021/acs.est.7b05559 (2018).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Cooper, D. A. & Corcoran, P. L. Effects of mechanical and chemical processes on the degradation of plastic beach debris on the island of Kauai, Hawaii. Mar. Pollut. Bull. 60(5), 650–654. https://doi.org/10.1016/j.marpolbul.2009.12.026 (2010).Article 
CAS 
PubMed 

Google Scholar 
O’Brine, T. & Thompson, R. C. Degradation of plastic carrier bags in the marine environment. Mar. Pollut. Bull. 60(12), 2279–2283. https://doi.org/10.1016/j.marpolbul.2010.08.005 (2010).Article 
CAS 
PubMed 

Google Scholar 
Song, Y. K. et al. Combined effects of UV exposure duration and mechanical abrasion on microplastic fragmentation by polymer type. Environ. Sci Technol. 51(8), 4368–4376. https://doi.org/10.1021/acs.est.6b06155 (2017).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Adobe Inc. (2019). Adobe illustrator. Retrieved from https://www.adobe.com/Products/Illustrator.Chamas, A. et al. Degradation rates of plastics in the environment. ACS Sustain. Chem. Eng. 8(9), 3494–3511. https://doi.org/10.1021/acssuschemeng.9b06635 (2020).Article 
CAS 

Google Scholar 
Councell, T. B., Duckenfield, K. U., Landa, E. R. & Callender, E. Tire-wear particles as a source of zinc to the environment. Environ. Sci. Technol. 38(15), 4206–4214. https://doi.org/10.1021/es034631f (2004).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Awet, T. T. et al. Effects of polystyrene nanoparticles on the microbiota and functional diversity of enzymes in soil. Environ. Sci. Eur. https://doi.org/10.1186/s12302-018-0140-6 (2018).Article 
PubMed 
PubMed Central 

Google Scholar 
Chung, H., Son, Y., Yoon, T. K., Kim, S. & Kim, W. The effect of multi-walled carbon nanotubes on soil microbial activity. Ecotoxicol. Environ. Saf. 74(4), 569–575. https://doi.org/10.1016/j.ecoenv.2011.01.004 (2011).Article 
CAS 
PubMed 

Google Scholar 
Huber, M., Welker, A. & Helmreich, B. Critical review of heavy metal pollution of traffic area runoff: Occurrence, influencing factors, and partitioning. Sci. Total Environ. 541, 895–919. https://doi.org/10.1016/j.scitotenv.2015.09.033 (2016).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Miandad, R., Barakat, M. A., Aburiazaiza, A. S., Rehan, M. & Nizami, A. S. Catalytic pyrolysis of plastic waste: A review. Process Saf. Environ. Prot. 102, 822–838. https://doi.org/10.1016/j.psep.2016.06.022 (2016).Article 
CAS 

Google Scholar 
Zhang, X., Li, H., Cao, Q., Jin, L. & Wang, F. Upgrading pyrolytic residue from waste tires to commercial carbon black. Waste Manag. Res. 36(5), 436–444. https://doi.org/10.1177/0734242X18764292 (2018).Article 
CAS 
PubMed 

Google Scholar 
Zhu, D., Li, G., Wang, H. T. & Duan, G. L. Effects of nano- or microplastic exposure combined with arsenic on soil bacterial, fungal, and protistan communities. Chemosphere 281, 130998. https://doi.org/10.1016/j.chemosphere.2021.130998 (2021).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Pathan, S. I. et al. Soil Pollution from micro-and nanoplastic debris: A hidden and unknown biohazard. Sustainability 12(18), 1–31. https://doi.org/10.3390/su12187255 (2020).Article 
CAS 

Google Scholar 
Rillig, M. C. & Bonkowski, M. Microplastic and soil protists: A call for research. Environ. Pollut. 241, 1128–1131. https://doi.org/10.1016/j.envpol.2018.04.147 (2018).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Zettler, E. R., Mincer, T. J. & Amaral-Zettler, L. A. Life in the “Plastisphere”: Microbial communities on plastic marine debris. Environ. Sci. Technol. 47(13), 7137–7146. https://doi.org/10.1021/es401288x (2013).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Burns, E. E. & Boxall, A. B. A. Microplastics in the aquatic environment: Evidence for or against adverse impacts and major knowledge gaps. Environ. Toxicol. Chem. 37(11), 2776–2796. https://doi.org/10.1002/etc.4268 (2018).Article 
CAS 
PubMed 

Google Scholar 
Bradney, L. et al. Particulate plastics as a vector for toxic trace-element uptake by aquatic and terrestrial organisms and human health risk. Environ. Int. 2019(131), 104937. https://doi.org/10.1016/j.envint.2019.104937 (2018).Article 
CAS 

Google Scholar 
Duis, K. & Coors, A. Microplastics in the Aquatic and Terrestrial Environment: Sources (with a Specific Focus on Personal Care Products), fate and effects. Environ. Sci. Eur. 28(1), 1–25. https://doi.org/10.1186/s12302-015-0069-y (2016).Article 
CAS 

Google Scholar 
Geyer, R., Jambeck, J. R. & Law, K. L. Production, use, and fate of all plastics ever made. Sci. Adv. 3(7), 25–29. https://doi.org/10.1126/sciadv.1700782 (2017).Article 
CAS 

Google Scholar 
Jayasiri, H. B., Purushothaman, C. S. & Vennila, A. Quantitative analysis of plastic debris on recreational beaches in Mumbai, India. Mar. Pollut. Bull. 77(1–2), 107–112. https://doi.org/10.1016/j.marpolbul.2013.10.024 (2013).Article 
CAS 
PubMed 

Google Scholar 
Lassen, C., Hansen, S. F., Magnusson, K., Hartmann, N. B., Rehne Jensen, P., Nielsen, T. G. & Brinch, A. Microplastics occurrence, effects and sources of releases (2015).Weithmann, N. et al. Organic fertilizer as a vehicle for the entry of microplastic into the environment. Sci. Adv. https://doi.org/10.1126/sciadv.aap8060 (2018).Article 
PubMed 
PubMed Central 

Google Scholar 
Hidalgo-Ruz, V., Gutow, L., Thompson, R. C. & Thiel, M. Microplastics in the marine environment: A review of the methods used for identification and quantification. Environ. Sci. Technol. 46(6), 3060–3075. https://doi.org/10.1021/es2031505 (2012).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Boenigk, J., Matz, C., Jürgens, K. & Arndt, H. Confusing selective feeding with differential digestion in bacterivorous nanoflagellates. J. Eukaryot. Microbiol. 48(4), 425–432. https://doi.org/10.1111/j.1550-7408.2001.tb00175.x (2001).Article 
CAS 
PubMed 

Google Scholar 
Boenigk, J., Matz, C., Jürgens, K. & Arndt, H. Food concentration-dependent regulation of food selectivity of interception-feeding bacterivorous nanoflagellates. Aquat. Microb. Ecol. 27(2), 195–202. https://doi.org/10.3354/ame027195 (2002).Article 

Google Scholar 
Wright, S. L., Thompson, R. C. & Galloway, T. S. The physical impacts of microplastics on marine organisms: A review. Environ. Pollut. 178, 483–492. https://doi.org/10.1016/j.envpol.2013.02.031 (2013).Article 
CAS 
PubMed 

Google Scholar 
Moore, C. J. Synthetic polymers in the marine environment: A rapidly increasing, long-term threat. Environ. Res. 108(2), 131–139. https://doi.org/10.1016/j.envres.2008.07.025 (2008).Article 
CAS 
PubMed 

Google Scholar 
Fu, S. F. et al. Exposure to polystyrene nanoplastic leads to inhibition of anaerobic digestion system. Sci. Total Environ. 625, 64–70. https://doi.org/10.1016/j.scitotenv.2017.12.158 (2018).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Bock, C. et al. Factors shaping community patterns of protists and bacteria on a European scale. Environ. Microbiol. 22(6), 2243–2260. https://doi.org/10.1111/1462-2920.14992 (2020).Article 
PubMed 

Google Scholar 
Besseling, E., Wang, B., Lürling, M. & Koelmans, A. A. Nanoplastic affects growth of S. obliquus and reproduction of D. magna. Environ. Sci. Technol. 48(20), 12336–12343. https://doi.org/10.1021/es503001d (2014).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Brown, D. M., Wilson, M. R., MacNee, W., Stone, V. & Donaldson, K. Size-dependent proinflammatory effects of ultrafine polystyrene particles: A role for surface area and oxidative stress in the enhanced activity of ultrafines. Toxicol. Appl. Pharmacol. 175(3), 191–199. https://doi.org/10.1006/taap.2001.9240 (2001).Article 
CAS 
PubMed 

Google Scholar 
Jeong, C. B. et al. Microplastic size-dependent toxicity, oxidative stress induction, and p-JNK and p-P38 activation in the monogonont rotifer (Brachionus Koreanus). Environ. Sci. Technol. 50(16), 8849–8857. https://doi.org/10.1021/acs.est.6b01441 (2016).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Kang, H. C., Jeong, H. J., Jang, S. H. & Lee, K. H. Feeding by common heterotrophic protists on the phototrophic dinoflagellate Biecheleriopsis adriatica (Suessiaceae) compared to that of other suessioid dinoflagellates. Algae 34(2), 127–140. https://doi.org/10.4490/algae.2019.34.5.29 (2019).Article 
CAS 

Google Scholar 
Sjollema, S. B., Redondo-Hasselerharm, P., Leslie, H. A., Kraak, M. H. S. & Vethaak, A. D. Do plastic particles affect microalgal photosynthesis and growth?. Aquat. Toxicol. 170, 259–261. https://doi.org/10.1016/j.aquatox.2015.12.002 (2016).Article 
CAS 
PubMed 

Google Scholar 
Rossi, G., Barnoud, J. & Monticelli, L. Polystyrene nanoparticles perturb lipid membranes. J. Phys. Chem. Lett. 5(1), 241–246. https://doi.org/10.1021/jz402234c (2014).Article 
CAS 
PubMed 

Google Scholar 
Brandts, I. et al. Effects of nanoplastics on mytilus galloprovincialis after individual and combined exposure with carbamazepine. Sci. Total Environ. 643, 775–784. https://doi.org/10.1016/j.scitotenv.2018.06.257 (2018).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Ciacci, C. et al. Nanoparticle-biological interactions in a marine benthic foraminifer. Sci. Rep. https://doi.org/10.1038/s41598-019-56037-2 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Kim, J. A. et al. Low dose of amino-modified nanoparticles induces cell cycle arrest. ACS Nano 7(9), 7483–7494 (2013).Article 
CAS 
PubMed 

Google Scholar 
Mao, Y. et al. Phytoplankton response to polystyrene microplastics: Perspective from an entire growth period. Chemosphere https://doi.org/10.1016/j.chemosphere.2018.05.170 (2018).Article 
PubMed 

Google Scholar 
Wang, F. et al. Time resolved study of cell death mechanisms induced by amine-modified polystyrene nanoparticles. Nanoscale 5(22), 10868–10876. https://doi.org/10.1039/c3nr03249c (2013).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Xia, T. et al. Comparison of the abilities of ambient and manufactured nanoparticles to induce cellular toxicity according to an oxidative stress paradigm. Nano Lett. 6(8), 1794–1807. https://doi.org/10.1021/nl061025k (2006).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Lagarde, F. et al. Microplastic interactions with freshwater microalgae: Hetero-aggregation and changes in plastic density appear strongly dependent on polymer type. Environ. Pollut. 215, 331–339. https://doi.org/10.1016/j.envpol.2016.05.006 (2016).Article 
CAS 
PubMed 

Google Scholar 
Bhattacharya, P., Lin, S., Turner, J. P. & Ke, P. C. Physical adsorption of charged plastic nanoparticles affects algal photosynthesis. J. Phys. Chem. C 114(39), 16556–16561. https://doi.org/10.1021/jp1054759 (2010).Article 
CAS 

Google Scholar 
Johansen, J. L., Rønn, R. & Ekelund, F. Toxicity of cadmium and zinc to small soil protists. Environ. Pollut. 242, 1510–1517. https://doi.org/10.1016/j.envpol.2018.08.034 (2018).Article 
CAS 
PubMed 

Google Scholar 
Díaz, S., Martín-González, A. & Carlos Gutiérrez, J. Evaluation of heavy metal acute toxicity and bioaccumulation in soil ciliated protozoa. Environ. Int. 32(6), 711–717. https://doi.org/10.1016/j.envint.2006.03.004 (2006).Article 
CAS 
PubMed 

Google Scholar 
Subba, P. et al. Zinc stress induces physiological, ultra-structural and biochemical changes in mandarin orange (Citrus Reticulata Blanco) seedlings. Physiol. Mol. Biol. Plants 20(4), 461–473. https://doi.org/10.1007/s12298-014-0254-2 (2014).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Corcoll, N. et al. The effect of metals on photosynthesis processes and diatom metrics of biofilm from a metal-contaminated river: A translocation experiment. Ecol. Indic. 18, 620–631. https://doi.org/10.1016/j.ecolind.2012.01.026 (2012).Article 
CAS 

Google Scholar 
Moffett, B. F. et al. Zinc contamination decreases the bacterial diversity of agricultural soil. FEMS Microbiol. Ecol. 43(1), 13–19. https://doi.org/10.1016/S0168-6496(02)00448-8 (2003).Article 
CAS 
PubMed 

Google Scholar 
Kuperman, R. G. & Carreiro, M. M. Soil heavy metal concentrations, microbial biomass and enzyme activities in a contaminated grassland ecosystem. Soil Biol. Biochem. 29(2), 179–190. https://doi.org/10.1016/S0038-0717(96)00297-0 (1997).Article 
CAS 

Google Scholar 
Masmoudi, S. et al. Cadmium, copper, sodium and zinc effects on diatoms: From heaven to hell-a review. Cryptogam Algol 34(2), 185–225. https://doi.org/10.7872/crya.v34.iss2.2013.185 (2013).Article 

Google Scholar 
Gadd, G. M. & de Rome, L. Biosorption of copper by fungal melanin. Appl. Microbiol. Biotechnol. 29(6), 610–617. https://doi.org/10.1007/BF00260993 (1988).Article 
CAS 

Google Scholar 
Khan, M. & Scullion, J. Effects of metal (Cd, Cu, Ni, Pb or Zn) enrichment of sewage-sludge on soil micro-organisms and their activities. Appl. Soil. Ecol. 20(2), 145–155. https://doi.org/10.1016/S0929-1393(02)00018-5 (2002).Article 

Google Scholar 
Guillard, R. R. L. & Lorenzen, C. J. Yellow-green algae with chlorophyllide C12. J. Phycol. 8(1), 10–14. https://doi.org/10.1111/j.1529-8817.1972.tb03995.x (1972).Article 
CAS 

Google Scholar 
Zagata, P., Kopańska, M., Greczek-Stachura, M. & Burnecki, T. Acute toxicity of metals: Nickel and zinc to Paramecium bursaria and its endosymbionts. J. Microbiol. Biotechnol. Food Sci. 04, 128–131. https://doi.org/10.15414/jmbfs.2015.4.special2.128-131 (2015).Article 
CAS 

Google Scholar 
Lenz, R., Enders, K. & Nielsen, T. G. Microplastic exposure studies should be environmentally realistic. Proc. Natl. Acad. Sci. U. S. A. 113(29), E4121–E4122. https://doi.org/10.1073/pnas.1606615113 (2016).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Schertzinger, G., Ruchter, N. & Sures, B. Metal accumulation in sediments and amphipods downstream of combined sewer overflows. Sci. Total Environ. 616–617, 1199–1207. https://doi.org/10.1016/j.scitotenv.2017.10.199 (2018).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Erasmus, J. H. et al. Metal accumulation in riverine macroinvertebrates from a platinum mining region. Sci. Total Environ. https://doi.org/10.1016/j.scitotenv.2019.134738 (2020).Article 
PubMed 

Google Scholar 
Pradhan, S., Hedberg, J., Blomberg, E., Wold, S. & Odnevall Wallinder, I. Effect of sonication on particle dispersion, administered dose and metal release of non-functionalized, non-inert metal nanoparticles. J. Nanopart. Res. 18(9), 1–14. https://doi.org/10.1007/s11051-016-3597-5 (2016).Article 
CAS 

Google Scholar 
Taurozzi, J. S., Hackley, V. A. & Wiesner, M. R. Preparation of nanoparticle dispersions from powdered material using ultrasonic disruption. NIST Spec. Publ. 1200–2, 1–15 (2012).
Google Scholar 
Graupner, N. et al. Effects of short-term flooding on aquatic and terrestrial microeukaryotic communities: A mesocosm approach. Aquat. Microb. Ecol. 80(3), 257–272. https://doi.org/10.3354/ame01853 (2017).Article 

Google Scholar 
Strasser, R., Srivastava, A. & Tsimilli-Michael, M. The fluorescence transient as a tool to characterize and screen photosynthetic samples. In Probing Photosynthesis Mechanisms, Regulation and Adaption (eds Yanus, M. et al.) (Taylor and Francis, 2020).
Google Scholar 
Thwe, A. & Kasemsap, P. Quantification of OJIP fluorescence transient in tomato plants under acute ozone stress (2015).Amaral-Zettler, L. A., McCliment, E. A., Ducklow, H. W. & Huse, S. M. A method for studying protistan diversity using massively parallel sequencing of V9 hypervariable regions of small-subunit ribosomal RNA genes. PLoS ONE 4(7), 1–9. https://doi.org/10.1371/journal.pone.0006372 (2009).Article 
CAS 

Google Scholar 
Medlin, L., Elwood, H. J., Stickel, S. & Sogin, M. L. The characterization of enzymatically amplified eukaryotic 16S-like RRNA-coding regions. Gene 71(2), 491–499. https://doi.org/10.1016/0378-1119(88)90066-2 (1988).Article 
CAS 
PubMed 

Google Scholar 
Andrews, S. FastQC: A quality control tool for high throughput sequence data (2015).Lange, A. et al. AmpliconDuo: A split-sample filtering protocol for high-throughput amplicon sequencing of microbial communities. PLoS ONE 10(11), 1–22. https://doi.org/10.1371/journal.pone.0141590 (2015).Article 
CAS 

Google Scholar 
Schmieder, R. & Edwards, R. Quality control and preprocessing of metagenomic datasets. Bioinformatics 27(6), 863–864. https://doi.org/10.1093/bioinformatics/btr026 (2011).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Masella, P. A., Bartram, A. K., Truszkowski, J. M., Brow, D. G. & Neufeld, J. D. PANDAseq: Paired-end assembler for illumina sequences. BMC Bioinform. https://doi.org/10.1186/1471-2105-13-31 (2012).Article 

Google Scholar 
Edgar, R. C., Haas, B. J., Clemente, J. C., Quince, C. & Knight, R. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27(16), 2194–2200. https://doi.org/10.1093/bioinformatics/btr381 (2011).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Mahé, F., Rognes, T., Quince, C., de Vargas, C. & Dunthorn, M. Swarm: Robust and fast clustering method for amplicon-based studies. PeerJ 2014(1), 1–13. https://doi.org/10.7717/peerj.593 (2014).Article 

Google Scholar 
Callahan, B. J. et al. DADA2: High-resolution sample inference from illumina amplicon data. Nat. Methods 13(7), 581–583. https://doi.org/10.1038/nmeth.3869 (2016).Article 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Welzel, M. et al. Natrix: A snakemake-based workflow for processing, clustering, and taxonomically assigning amplicon sequencing reads. BMC Bioinform. 21(1), 1–14. https://doi.org/10.1186/s12859-020-03852-4 (2020).Article 
CAS 

Google Scholar 
Oksanen, J. Package “vegan” Title Community Ecology Package (2022).R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.r-project.org/.Chen, W., Simpson, J. & Leveque, C. RAM: R for amplicon-sequencing-based microbial-ecology (2018).Yarza, P. et al. Uniting the classification of cultured and uncultured bacteria and archaea using 16S RRNA gene sequences. Nat. Rev. Microbiol. 12(9), 635–645. https://doi.org/10.1038/nrmicro3330 (2014).Article 
CAS 
PubMed 

Google Scholar 
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15(12), 1–21. https://doi.org/10.1186/s13059-014-0550-8 (2014).Article 
CAS 

Google Scholar 
Palarea-Albaladejo, J. & Martín-Fernández, J. A. ZCompositions—R package for multivariate imputation of left-censored data under a compositional approach. Chemom. Intell. Lab. Syst. 143, 85–96. https://doi.org/10.1016/j.chemolab.2015.02.019 (2015).Article 
CAS 

Google Scholar 
Gloor, G. B., Macklaim, J. M., Pawlowsky-Glahn, V. & Egozcue, J. J. Microbiome datasets are compositional: And this is not optional. Front. Microbiol. 8, 1–6. https://doi.org/10.3389/fmicb.2017.02224 (2017).Article 

Google Scholar 
Dusaucy, J., Gateuille, D., Perrette, Y. & Naffrechoux, E. Microplastic pollution of worldwide lakes. Environ. Pollut. 284, 117075. https://doi.org/10.1016/j.envpol.2021.117075 (2021).Article 
CAS 
PubMed 

Google Scholar 
Vardhan, K. H., Kumar, P. S. & Panda, R. C. A review on heavy metal pollution, toxicity and remedial measures: Current trends and future perspectives. J. Mol. Liq. 290, 111197. https://doi.org/10.1016/j.molliq.2019.111197 (2019).Article 
CAS 

Google Scholar 
Damare, V. S. Diversity of thraustochytrid protists isolated from brown alga, Sargassum cinereum using 18S RDNA sequencing and their morphological response to heavy metals. J. Mar. Biol. Assoc. 95(2), 265–276. https://doi.org/10.1017/S0025315414001696 (2015).Article 
CAS 

Google Scholar 
Giongo, A. et al. Adaption of Microbial communities to the hostile environment in the Doce river after the collapse of two iron ore tailing dams. Heliyon https://doi.org/10.1016/j.heliyon.2020.e04778 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Kelly, J. J., Häggblom, M. M. & Tate, R. L. Effects of heavy metal contamination and remediation on soil microbial communities in the vicinity of a zinc smelter as indicated by analysis of microbial community phospholipid fatty acid profiles. Biol. Fertil. Soils 38(2), 65–71. https://doi.org/10.1007/s00374-003-0642-1 (2003).Article 
CAS 

Google Scholar 
Baddar, Z. E., Peck, E. & Xu, X. Temporal deposition of copper and zinc in the sediments of metal removal constructed wetlands. PLoS ONE 16, 1–14. https://doi.org/10.1371/journal.pone.0255527 (2021).Article 
CAS 

Google Scholar 
Li, X., Shen, Z., Wai, O. W. H. & Li, Y. S. Chemical partitioning of heavy metal contaminants in sediments of the Pearl River Estuary. Chem. Speciat. Bioavailab. 12(1), 17–25. https://doi.org/10.3184/095422900782775607 (2000).Article 
CAS 

Google Scholar 
Müller, B. & Sigg, L. Interaction of trace metals with natural particle surfaces: Comparison between adsorption experiments and field measurements—Dedicated to Werner Stumm for his 65th birthday. Aquat. Sci. 52(1), 75–92. https://doi.org/10.1007/BF00878242 (1990).Article 

Google Scholar 
Bradl, H. B. Adsorption of heavy metal ions on soils and soils constituents. J. Colloid Interface Sci. 277(1), 1–18. https://doi.org/10.1016/j.jcis.2004.04.005 (2004).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Siegel, F. R. Environmental Geochemistry of Potentially Toxic Heavy Metals (Springer-Verlag, 2002).Book 

Google Scholar 
Vig, K., Megharaj, M., Sethunathan, N. & Naidu, R. Bioavailability and toxicity of cadmium to microorganisms and their activities in soil: A review. Adv. Environ. Res. 8(1), 121–135. https://doi.org/10.1016/S1093-0191(02)00135-1 (2003).Article 
CAS 

Google Scholar 
Nicolau, A., Mota, M. & Lima, N. Physiological responses of tetrahymena pyriformis to copper, zinc, cycloheximide and triton X-100. FEMS Microbiol. Ecol. 30(3), 209–216. https://doi.org/10.1016/S0168-6496(99)00057-4 (1999).Article 
CAS 
PubMed 

Google Scholar 
Admiraal, W. et al. Short-term toxicity of zinc to microbenthic algae and bacteria in a metal polluted stream. Water Res. 33(9), 1989–1996. https://doi.org/10.1016/S0043-1354(98)00426-6 (1999).Article 
CAS 

Google Scholar 
Bradac, P., Navarro, E., Odzak, N., Behra, R. & Sigg, L. Kinetics of cadmium accumulation in periphyton under freshwater conditions. Environ. Toxicol. Chem. 28(10), 2108–2116. https://doi.org/10.1897/08-511R1.1 (2009).Article 
CAS 
PubMed 

Google Scholar 
Collard, J. & Matagne, R. F. Cd2+ resistance in wild-type and mutant strains of Chlamydomonas reinhardtii. Environ. Exp. Bot. 34(2), 235–244 (1994).Article 
CAS 

Google Scholar 
Wright, R. J., Gibson, M. I. & Christie-Oleza, J. A. Understanding microbial community dynamics to improve optimal microbiome selection. Microbiome 7(1), 1–14. https://doi.org/10.1186/s40168-019-0702-x (2019).Article 

Google Scholar 
Buffle, J. The key role of environmental colloids/nanoparticles for the sustainability of life. Environ. Chem. 3(3), 155–158. https://doi.org/10.1071/ENv3n3_ES (2006).Article 
CAS 

Google Scholar 
Nowack, B. & Bucheli, T. D. Occurrence, behavior and effects of nanoparticles in the environment. Environ. Pollut. 150(1), 5–22. https://doi.org/10.1016/j.envpol.2007.06.006 (2007).Article 
CAS 
PubMed 

Google Scholar 
Fetzer, I. et al. The extent of functional redundancy changes as species’ roles shift in different environments. Proc. Natl. Acad. Sci. U. S. A. 112(48), 14888–14893. https://doi.org/10.1073/pnas.1505587112 (2015).Article 
ADS 
CAS 
PubMed 
PubMed Central 

Google Scholar 
Biggs, C. R. et al. Does functional redundancy affect ecological stability and resilience? A review and meta-analysis. Ecosphere https://doi.org/10.1002/ecs2.3184 (2020).Article 

Google Scholar 
Fleeger, J. W. How do indirect effects of contaminants inform ecotoxicology? A review. Processes https://doi.org/10.3390/pr8121659 (2020).Article 

Google Scholar 
Oriekhova, O. & Stoll, S. Heteroaggregation of nanoplastic particles in the presence of inorganic colloids and natural organic matter. Environ. Sci. Nano. 5(3), 792–799. https://doi.org/10.1039/c7en01119a (2018).Article 
CAS 

Google Scholar 
Rowenczyk, L. et al. Heteroaggregates of polystyrene nanospheres and organic matter: Preparation, characterization and evaluation of their toxicity to algae in environmentally relevant conditions. Nanomaterials 11(2), 1–15. https://doi.org/10.3390/nano11020482 (2021).Article 
CAS 

Google Scholar 
Saavedra, J., Stoll, S. & Slaveykova, V. I. Influence of nanoplastic surface charge on eco-corona formation, aggregation and toxicity to freshwater zooplankton. Environ. Pollut. 252, 715–722. https://doi.org/10.1016/j.envpol.2019.05.135 (2019).Article 
CAS 
PubMed 

Google Scholar 
Bižic-Ionescu, M., Ionescu, D. & Grossart, H. P. Organic particles: Heterogeneous hubs for microbial interactions in aquatic ecosystems. Front. Microbiol. 9, 1–15. https://doi.org/10.3389/fmicb.2018.02569 (2018).Article 

Google Scholar 
Lespes, G., Faucher, S. & Slaveykova, V. I. natural nanoparticles, anthropogenic nanoparticles, where is the Frontier?. Front. Environ. Sci. 8, 1–5. https://doi.org/10.3389/fenvs.2020.00071 (2020).Article 

Google Scholar 
Stabnikova, O. et al. Microbial life on the surface of microplastics in natural waters. Appl. Sci. 11(24), 1–19. https://doi.org/10.3390/app112411692 (2021).Article 
CAS 

Google Scholar 
Suominen, S., Doorenspleet, K., Sinninghe Damsté, J. S. & Villanueva, L. Microbial community development on model particles in the deep sulfidic waters of the Black Sea. Environ. Microbiol. 23(6), 2729–2746. https://doi.org/10.1111/1462-2920.15024 (2021).Article 
CAS 
PubMed 

Google Scholar 
Wagner, S., Gondikas, A., Neubauer, E., Hofmann, T. & von der Kammer, F. Spot the difference: Engineered and natural nanoparticles in the environment-release, behavior, and fate. Angew. Chem. Int. Ed. 53(46), 12398–12419. https://doi.org/10.1002/anie.201405050 (2014).Article 
CAS 

Google Scholar 
Amelia, T. S. et al. Marine microplastics as vectors of major ocean pollutants and its hazards to the marine ecosystem and humans. Prog. Earth Planet. Sci. https://doi.org/10.1186/s40645-020-00405-4 (2021).Article 

Google Scholar 
Liu, J., Huang, J. & Che, F. Microalgae as feedstocks for biodiesel production. In Biodiesel—Feedstocks and Processing Technologies (ed. Stoytcheva, M.) (InTech, 2011). https://doi.org/10.5772/25600.Chapter 

Google Scholar 
Takamura, N., Kasai, F. & Watanabe, M. M. Effects of Cu, Cd and Zn on photosynthesis of freshwater benthic algae. J. Appl. Phycol. 1(1), 39–52. https://doi.org/10.1007/BF00003534 (1989).Article 
CAS 

Google Scholar 
Brembu, T., Jørstad, M., Winge, P., Valle, K. C. & Bones, A. M. Genome-wide profiling of responses to cadmium in the diatom Phaeodactylum tricornutum. Environ. Sci. Technol. 45(18), 7640–7647. https://doi.org/10.1021/es2002259 (2011).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Fernandez, J. C. & Henriques, F. S. Biochemical, physiological and structural effects of excess copper in plants. Bot. Rev. 57(3), 246–273 (1991).Article 

Google Scholar 
Haq, R. U., Rehman, A. & Shakoori, A. R. Effect of dichromate on population and growth of various protozoa isolated from industrial effluents. Folia Microbiol. 45(3), 275–278. https://doi.org/10.1007/bf02908959 (2000).Article 
CAS 

Google Scholar 
Rehman, A., Shakoori, F. R. & Shakoori, A. R. Heavy metal resistant freshwater ciliate, Euplotes mutabilis, isolated from industrial effluents has potential to decontaminate wastewater of toxic metals. Bioresour. Technol. 99(9), 3890–3895. https://doi.org/10.1016/j.biortech.2007.08.007 (2008).Article 
CAS 
PubMed 

Google Scholar 
Rehman, A., Ashraf, S., Qazi, J. I. & Shakoori, A. R. Uptake of lead by a ciliate, stylonychia mytilus, isolated from industrial effluents: Potential use in bioremediation of wastewater. Bull. Environ. Contam. Toxicol. 75(2), 290–296. https://doi.org/10.1007/s00128-005-0751-7 (2005).Article 
CAS 
PubMed 

Google Scholar 
Shakoori, A. R., Rehman, A. & ul-Haq, R. Multiple metal resistance in the ciliate protozoan, vorticella microstoma, isolated from industrial effluents and its potential in bioremediation of toxic wastes. Bull. Environ. Contam. Toxicol. 72(5), 1046–1051. https://doi.org/10.1007/s00128-004-0349-5 (2004).Article 
CAS 
PubMed 

Google Scholar 
Falasco, E. et al. Morphological abnormalities of diatom silica walls in relation to heavy metal contamination and artificial growth conditions. Water SA 35(5), 595–606. https://doi.org/10.4314/wsa.v35i5.49185 (2009).Article 
CAS 

Google Scholar 
Tadros, M. G., Mbuthia, P. & Smith, W. Differential response of marine diatoms to trace metals. Bull. Environ. Contam. Toxicol. 44(6), 826–831. https://doi.org/10.1007/BF01702170 (1990).Article 
CAS 
PubMed 

Google Scholar 
Wanner, M. et al. Soil testate amoebae and diatoms as bioindicators of an old heavy metal contaminated floodplain in Japan. Microb. Ecol. 79(1), 123–133. https://doi.org/10.1007/s00248-019-01383-x (2020).Article 
CAS 
PubMed 

Google Scholar 
Shi, J., Podola, B. & Melkonian, M. Application of a prototype-scale twin-layer photobioreactor for effective N and P removal from different process stages of municipal wastewater by immobilized microalgae. Bioresour. Technol. 154, 260–266. https://doi.org/10.1016/j.biortech.2013.11.100 (2014).Article 
CAS 
PubMed 

Google Scholar 
Li, T., Lin, G., Podola, B. & Melkonian, M. Continuous removal of zinc from wastewater and mine dump leachate by a microalgal biofilm PSBR. J. Hazard. Mater. 297, 112–118. https://doi.org/10.1016/j.jhazmat.2015.04.080 (2015).Article 
CAS 
PubMed 

Google Scholar 
Bruning, K. Infection of the diatom Asterionella by a chytrid. I. Effects of light on reproduction and infectivity of the parasite. J. Plankton Res. 13(1), 103–117. https://doi.org/10.1093/plankt/13.1.103 (1991).Article 

Google Scholar 
Carney, L. T. & Lane, T. W. Parasites in algae mass culture. Front. Microbiol. 5, 1–8. https://doi.org/10.3389/fmicb.2014.00278 (2014).Article 

Google Scholar 
Hanic, L. A., Sekimoto, S. & Bates, S. S. Oomycete and chytrid infections of the marine diatom Pseudo-nitzschia pungens (Bacillariophyceae) from Prince Edward Island. Botany 87(11), 1096–1105. https://doi.org/10.1139/B09-070 (2009).Article 
CAS 

Google Scholar 
Sun, A. et al. Fertilization alters protistan consumers and parasites in crop-associated microbiomes. Environ. Microbiol. 23(4), 2169–2183. https://doi.org/10.1111/1462-2920.15385 (2021).Article 
CAS 
PubMed 

Google Scholar 
Scholz, B., Guillou, L., Marano, A. V., Neuhauser, S. & Brooke, K. Europe PMC funders group zoosporic parasites infecting marine diatoms—A black box that needs to be opened. Fungal Ecol. https://doi.org/10.1016/j.funeco.2015.09.002.Zoosporic (2017).Article 

Google Scholar 
Peacock, E. E., Olson, R. J. & Sosik, H. M. Parasitic infection of the diatom Guinardia delicatula, a recurrent and ecologically important phenomenon on the New England Shelf. Mar. Ecol. Prog. Ser. 503, 1–10. https://doi.org/10.3354/meps10784 (2014).Article 
ADS 

Google Scholar 
Duarte, S., Pascoal, C. & Cássio, F. Effects of zinc on leaf decomposition by fungi in streams: Studies in microcosms. Microb. Ecol. 48(3), 366–374. https://doi.org/10.1007/s00248-003-2032-5 (2004).Article 
CAS 
PubMed 

Google Scholar 
Kammerlander, B. et al. High diversity of protistan plankton communities in remote high mountain lakes in the European Alps and the Himalayan Mountains. FEMS Microbiol. Ecol. 91(4), 1–10. https://doi.org/10.1093/femsec/fiv010 (2015).Article 
CAS 

Google Scholar 
Sieber, G., Beisser, D., Bock, C. & Boenigk, J. Protistan and fungal diversity in soils and freshwater lakes are substantially different. Sci. Rep. 10(1), 1–11. https://doi.org/10.1038/s41598-020-77045-7 (2020).Article 
CAS 

Google Scholar 
Gunaalan, K., Fabbri, E. & Capolupo, M. The hidden threat of plastic leachates: A critical review on their impacts on aquatic organisms. Water Res. https://doi.org/10.1016/j.watres.2020.116170 (2020).Article 
PubMed 

Google Scholar 
Tetu, S. G., Sarker, I., Schrameyer, V., Pickford, R., Elbourne, L. D., Moore, L.R. & Paulsen, I.T. Plastic leachates impair growth and oxygen production in Prochlorococcus, the ocean’s most abundant photosynthetic bacteria. Commun. Biol. 2(1), 1–9. https://doi.org/10.1038/s42003-019-0410-x (2019).Gouin, T., Roche, N., Lohmann, R. & Hodges, G. A Thermodynamic approach for assessing the environmental exposure of chemicals absorbed to microplastic. Environ. Sci. Technol. 45(4), 1466–1472. https://doi.org/10.1021/es1032025 (2011).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Lohmann, R. Microplastics are not important for the cycling and bioaccumulation of organic pollutants in the oceans—But should microplastics be considered POPs themselves?. Integr. Environ. Assess. Manag. 13(3), 460–465. https://doi.org/10.1002/ieam.1914 (2017).Article 
CAS 
PubMed 

Google Scholar 
Sukkasem, C. & Laehlah, S. An economical upflow bio-filter circuit (UBFC): A biocatalyst microbial fuel cell for sulfate-sulfide rich wastewater treatment. Environ. Sci. 1(2), 161–168. https://doi.org/10.1039/c4ew00028e (2015).Article 
CAS 

Google Scholar 
Abatenh, E., Gizaw, B., Tsegaye, Z. & Wassie, M. The role of microorganisms in bioremediation-A review. Open J. Environ. Biol. 2(1), 38–46. https://doi.org/10.17352/ojeb (2017).Article 

Google Scholar 
Zrimec, J., Kokina, M., Jonasson, S., Zorrilla, F. & Zelezniak, A. Plastic-degrading potential across the global microbiome correlates with recent pollution trends. MBio https://doi.org/10.1128/mBio (2021).Article 
PubMed 
PubMed Central 

Google Scholar 
Siver, P. A. Synurophyte algae. In Freshwater Algae of North America. Ecology and classification (eds Wehr, J. D. & Sheath, R. G.) 523–558 (Elsevier, 2003).Chapter 

Google Scholar 
Andersen, R. A. Molecular systematics of the chrysophyceae and synurophyceae. In Unravelling the Algae: The Past, Present, and Future of Algal Systematics (eds Brodie, J. & Lewis, J.) 285–314 (CRC Press, Boca Raton, 2007).Chapter 

Google Scholar 
Engin, I. K., Cekmecelioglu, D., Yücel, A. M. & Oktem, H. A. Evaluation of heterotrophic and mixotrophic cultivation of novel Micractinium Sp. ME05 on vinasse and its scale up for biodiesel production. Bioresour. Technol. 251, 128–134. https://doi.org/10.1016/j.biortech.2017.12.023 (2018).Article 
CAS 
PubMed 

Google Scholar 
Patrick, R. Ecology of freshwater diatoms and diatom communities. In The Biology of Diatoms (ed. Werner, D.) 284–332 (University of California Press, 1977).
Google Scholar 
Findenig, B. M., Chatzinotas, A. & Boenigk, J. Taxonomic and ecological characterization of stomatocysts of spumella-like flagellates (Chrysophyceae). J. Phycol. 46(5), 868–881. https://doi.org/10.1111/j.1529-8817.2010.00892.x (2010).Article 

Google Scholar 
Perez-Garcia, O., Escalante, F. M. E., de-Bashan, L. E. & Bashan, Y. Heterotrophic cultures of microalgae: Metabolism and potential products. Water Res. 45(1), 11–36. https://doi.org/10.1016/j.watres.2010.08.037 (2011).Article 
CAS 
PubMed 

Google Scholar 
Preisig, H. R. & Hibberd, D. J. Ultrastructure and taxonomy of Paraphysomonas (Chrysophyceae) and related genera 3. Nord. J. Bot. 3(6), 695–723. https://doi.org/10.1111/j.1756-1051.1983.tb01481.x (1983).Article 

Google Scholar 
Atkins, M. S. et al. Tolerance of flagellated protists to high sulfide and metal concentrations potentially encountered at deep-sea hydrothermal vents. Mar. Ecol. Prog. Ser. 226, 63–75. https://doi.org/10.3354/meps226063 (2002).Article 
ADS 
CAS 

Google Scholar 
Manru, G., Weisong, F. & Yunfen, S. Ecological study on protozoa in the sediment of the three-gorges area of the Changjiang River. Chin. J. Oceanol. Limnol. 6(3), 272–280. https://doi.org/10.1007/BF02846505 (1988).Article 

Google Scholar 
Tomilina, I. I., Gremyachikh, V. A., Myl’Nikov, A. P. & Komov, V. T. The effect of metal oxide nanoparticles (CeO2, TiO2, and ZnO) on biological parameters of freshwater nanoflagellates and crustaceans. Dokl. Biol. Sci. 436(1), 53–55. https://doi.org/10.1134/S0012496611010169 (2011).Article 
CAS 
PubMed 

Google Scholar 
Schampera, C. et al. Exposure to nanoplastics affects the outcome of infectious disease in phytoplankton. Environ. Pollut. https://doi.org/10.1016/j.envpol.2021.116781 (2021).Article 
PubMed 

Google Scholar 
Gonçalves, J. M., Sousa, V. S., Teixeira, M. R. & Bebianno, M. J. Chronic toxicity of polystyrene nanoparticles in the marine mussel Mytilus galloprovincialis. Chemosphere https://doi.org/10.1016/j.chemosphere.2021.132356 (2021).Article 
PubMed 

Google Scholar 
Kelpsiene, E., Torstensson, O., Ekvall, M. T., Hansson, L. A. & Cedervall, T. Long-term exposure to nanoplastics reduces life-time in Daphnia magna. Sci. Rep. 10(1), 1–7. https://doi.org/10.1038/s41598-020-63028-1 (2020).Article 
CAS 

Google Scholar 
Amin, N. M. Techniques for assessment of heavy metal toxicity using Acanthamoeba Sp, a small, naked and free-living amoeba. Funct. Ecosyst. https://doi.org/10.5772/36008 (2012).Article 

Google Scholar 
Amin, N. M., Azhar, N. & Shazili, M. Cytotoxic effects of mercury, cadmium, lead and zinc on Acanthamoeba Castellanii (2006).Gnecco, I., Berretta, C., Lanza, L. G. & la Barbera, P. Storm water pollution in the urban environment of Genoa, Italy. Atmos. Res. 77, 60–73. https://doi.org/10.1016/j.atmosres.2004.10.017 (2005).Article 
CAS 

Google Scholar 
Heim, R. R. An overview of weather and climate extremes—Products and trends. Weather Clim. Extrem. 10, 1–9. https://doi.org/10.1016/j.wace.2015.11.001 (2015).Article 

Google Scholar 
Saiki, M. K., Castleberry, D. T., May, T. W., Martin, B. A. & Bullard, F. N. Copper, cadmium, and zinc concentrations in aquatic food chains from the upper Sacramento River (California) and selected tributaries. Arch. Environ. Contam. Toxicol. 29(4), 484–491. https://doi.org/10.1007/BF00208378 (1995).Article 
CAS 

Google Scholar 
Wagner, S. et al. Tire wear particles in the aquatic environment—A review on generation, analysis, occurrence, fate and effects. Water Res. 139, 83–100. https://doi.org/10.1016/j.watres.2018.03.051 (2018).Article 
CAS 
PubMed 

Google Scholar 
Zhang, L., Zhao, B., Xu, G. & Guan, Y. Characterizing fluvial heavy metal pollutions under different rainfall conditions: Implication for aquatic environment protection. Sci. Total Environ. 635, 1495–1506. https://doi.org/10.1016/j.scitotenv.2018.04.211 (2018).Article 
ADS 
CAS 
PubMed 

Google Scholar 
Zhao, B. et al. Characterization of nitrosamines and nitrosamine precursors as non-point source pollutants during heavy rainfall events in an urban water environment. J. Hazard. Mater. 424, 127552. https://doi.org/10.1016/j.jhazmat.2021.127552 (2022).Article 
CAS 
PubMed 

Google Scholar 
Hüffer, T., Wagner, S., Reemtsma, T. & Hofmann, T. Sorption of organic substances to tire wear materials: Similarities and differences with other types of microplastic. TrAC Trends Anal. Chem. 113, 392–401. https://doi.org/10.1016/j.trac.2018.11.029 (2019).Article 
CAS 

Google Scholar 
Tamis, J. E. et al. Environmental risks of car tire microplastic particles and other road runoff pollutants. Microplastics Nanoplastics 1(1), 1–17. https://doi.org/10.1186/s43591-021-00008-w (2021).Article 

Google Scholar 
Chèvre, N. et al. Substance flow analysis as a tool for urban water management. Water Sci. Technol. 63(7), 1341–1348. https://doi.org/10.2166/wst.2011.132 (2011).Article 
PubMed 

Google Scholar 
Šourková, M., Adamcová, D. & Vaverková, M. D. The influence of microplastics from ground tyres on the acute, subchronical toxicity and microbial respiration of soil. Environ. MDPI 8(11), 1–14. https://doi.org/10.3390/environments8110128 (2021).Article 

Google Scholar 
Ye, G., Zhang, X., Yan, C., Lin, Y. & Huang, Q. Polystyrene microplastics induce microbial dysbiosis and dysfunction in surrounding seawater. Environ. Int. 156, 106724. https://doi.org/10.1016/j.envint.2021.106724 (2021).Article 
CAS 
PubMed 

Google Scholar  More

Each of the 23 key variables can be used for analysis. To validate the dataset, we used five plant-related variables (diversity of order, family, genus, species and species endemic to China) to demonstrate the process of using the dataset for analysis as follows:(1) For the four variables of plant taxa “order”, “family”, “genus” and “species”, the similarity and difference in spatial distribution pattern of diversity of different taxa in the Qinling-Daba Mountains climate transition zone were analyzed. The spatial distribution pattern of the diversity of the four taxa is shown in Fig. 3, which is increasingly lower from south (low latitude) to north (high latitude). This result is consistent with the classical latitudinal gradient model of plant diversity. The boundary between higher diversity in the south and lower diversity in the north is roughly located in the area of Funiu Mountains in the eastern Qinling-Daba Mountains, Taibai Mountains in the central Qinling-Daba Mountains and Baishui River in the western Qinling-Daba Mountains. However, with the reduction in taxon scale, the spatial distribution pattern of diversity tends to be complex. Orders (Fig. 3a) and families (Fig. 3b) can be divided by lines, while genera (Fig. 3c) need thicker lines, and species (Fig. 3d) can only be divided by polygons. Figure 3 shows that the taxonomic groups of families are more clearly divided, while species can only be divided by staggered bands. Therefore, when dividing the north–south boundary, the family taxon scale is appropriate, whereas the species scale is more appropriate when studying the north–south transition zone.Fig. 3Spatial distribution of diversity of orders, families, genera and species. (a) The blue dotted line is basically the dividing line of the order diversity of 50 species. The order diversity to the north of the blue dotted line is lower than 50 species, and the order diversity to the south of the blue dotted line is higher than 50 species. (b) The blue dotted line is basically the dividing line of the family diversity of 150 species. The family diversity to the north of the blue dotted line is lower than 150 species, and the family diversity to the south of the blue dotted line is higher than 150 species. (c) The thicker blue dotted line is basically the dividing line of genus diversity of 578–681 species. The genus diversity to the north of the blue dotted line is lower than 578 species, and the genus diversity to the south of the blue dotted line is higher than 681 species. (d) The blue area is basically the dividing line of species diversity of 1385–1618 species. The species diversity to the north of the blue dotted line is lower than 1385 species, and the species diversity to the south of the blue dotted line is higher than 1618 species.Full size imageThe dataset can also count the orders, families and genera that appear in 58 nature reserves, indicating that these orders, families and genera are widely distributed in this area, while the orders, families and genera that only appear in a single nature reserve indicate that these taxa are unique to this nature reserve in this area, reflecting their locality and uniqueness, which is helpful to understanding the specific distribution of plants in detail. The relevant statistics are as follows:
There are 28 orders present in every nature reserve:
Liliales, Dipsacales, Lamiales, Fabales, Ericales, Poales, Saxifragales, Malpighiales, Malvales, Asterales, Fagales, Gentianales, Geraniales, Ranunculales, Rosales, Solanales, Apiales, Cornales, Brassicales, Caryophyllales, Dioscoreales, Santalales, Myrtales, Asparagales, Celastrales, Sapindales, Alismatales, and Boraginales.The order that only appears in one nature reserve is Petrosaviales, which appears in the Dabashan Nature Reserve in Chongqing.
There are 51 families present in every nature reserve:
Liliaceae, Primulaceae, Plantaginaceae, Lamiaceae, Euphorbiaceae, Cannabaceae, Juncaceae, Fabaceae, Poaceae, Elaeagnaceae, Betulaceae, Apocynaceae, Violaceae, Malvaceae, Crassulaceae, Campanulaceae, Asteraceae, Orchidaceae, Polygonaceae, Orobanchaceae, Onagraceae, Gentianaceae, Geraniaceae, Ranunculaceae, Rubiaceae, Rosaceae, Caprifoliaceae, Thymelaeaceae, Apiaceae, Cyperaceae, Cornaceae, Paeoniaceae, Brassicaceae, Amaryllidaceae, Caryophyllaceae, Rhamnaceae, Santalaceae, Asparagaceae, Celastraceae, Sapindaceae, Adoxaceae, Araliaceae, Berberidaceae, Hydrangeaceae, Scrophulariaceae, Convolvulaceae, Urticaceae, Salicaceae, Papaveraceae, Iridaceae, and Boraginaceae.There are 15 families that only appear in one nature reserve, as shown in Table 2.Table 2 Endemic families of the nature reserves in the Qinling-Daba Mountains and surrounding areas.Full size table
There are 54 genera present in every nature reserve:
Patrinia, Polygonum, Sanicula, Plantago, Allium, Delphinium, Euphorbia, Juncus, Cynanchum, Trigonotis, Artemisia, Sorbus, Polygonatum, Scutellaria, Cirsium, Viburnum, Ajuga, Viola, Galium, Geranium, Salix, Epilobium, Gentiana, Ranunculus, Malus, Acer, Rubia, Rosa, Torilis, Lonicera, Adenophora, Philadelphus, Cornus, Paeonia, Rhamnus, Rumex, Carex, Thalictrum, Asparagus, Carpesium, Clematis, Potentilla, Euonymus, Eleutherococcus, Berberis, Spiraea, Rubus, Populus, Vicia, Silene, Iris, Poa, Aster, and Buddleja.There were 225 genera that only appeared in one nature reserve, as shown in Figshare file 269.(2) For the “species endemic to China” variable of plants, we can see from the diversity distribution pattern of species endemic to China in this region (Fig. 4) that the number of endemic species in the Qinling-Daba Mountains is higher than that of species outside of the region, which reflects the strong transition zone in the Qinling-Daba Mountains. The variables of species endemic to China obtained from the Qinling-Daba Mountains and their surroundings were clustered by the Bray–Curtis dissimilarity measure70 and Ward’s minimum variance (the clustering method recommended for plant cluster analysis). The clustering results are shown in Fig. 5a. At the same time, the clustering results are displayed in space. Figure 5b shows that category 3 extends from the east outside the Qinling-Daba Mountains to the Baishuijiang Nature Reserve inside the western Qinling-Daba Mountains, which is consistent with the fact that the Qinling-Daba Mountains are an important ecogeographical “corridor” connecting the east and the west.Fig. 4Spatial distribution of diversity of species endemic to China in the Qinling-Daba Mountains and adjacent areas.Full size imageFig. 5(a) Clustering results of Ward’s connection aggregation of species endemic to China in 58 nature reserves. (b) Spatial distribution of clustering results of species endemic to China; the larger the dot and the darker the color, the earlier it is merged into this category, and the smaller the dot and the lighter the color, the later it is merged into this category.Full size image More