{"paper_id":"1c087a21-28af-4e7b-98b2-e37ca3687878","body_text":"1\n1 Title: Biomonitoring in the Anthropocene: Urban estuary environmental DNA tracks\n2 marine fish, terrestrial wildlife, and human diet\n3\n4\n5\n6 Mark Y. Stoeckle 1, Jesse H. Ausubel1\n7\n8 1Program for the Human Environment, The Rockefeller University, New York, New York, \n9 United States of America\n10\n11 *Corresponding author\n12 Email: mark.stoeckle@rockefeller.edu (MYS)\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n2\n13 Abstract\n14 Managing human impacts in urban estuaries asks for up-to-date monitoring of marine \n15 life. Here we analyze vertebrate eDNA in New York City’s East River, a rocky estuary channel \n16 difficult to survey with mechanical gear and subject to wastewater discharge. We collected water \n17 samples weekly for one year and applied spike-in metabarcoding to quantify vertebrate eDNA. \n18 Replication experiments demonstrated good reproducibility above about 10 eDNA copies/PCR. \n19 We propose a fish censusing scale based on absolute eDNA abundance. Local marine fish eDNA \n20 followed a classic hollow curve species abundance distribution over four orders of magnitude, \n21 with abundant and common taxa comprising about 25% of species and 95% of fish eDNA. There \n22 was a 10-fold increase in local marine fish eDNA in summer and seasonal differences among \n23 taxa consistent with known phenology. Two fish species were newly abundant in comparison to \n24 an eDNA survey at the same site in 2016. Levels of other vertebrate eDNA—domesticated \n25 animal, non-fish wildlife, and dietary fish—were correlated with human eDNA levels, consistent \n26 with a shared wastewater source. Wastewater eDNA identified the commonest urban mammals, \n27 land birds, and household pets. Proportions of dietary animal eDNA in wastewater closely \n28 approximated proportions in national consumption statistics, opening a window into human diet \n29 assessment. Effort and cost for the weekly shoreline survey were modest. Vertebrate eDNA \n30 metabarcoding with spike-in quantification enabled weekly monitoring of urban estuary fish \n31 populations, identified overlooked newly abundant species, and reported on terrestrial wildlife \n32 and human diet.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n3\n33 Introduction\n34 Many estuaries are sites of major cities and at the same time essential habitat for wildlife. \n35 An urban estuary represents a profound transformation of the physical and living landscape \n36 [1,2]. Hazards to the aquatic biosphere may arise from dredging, commercial and residential \n37 wastewater discharge, maritime traffic, ballast water disposal, and engineering modifications \n38 such as dikes, piers, seawalls, and filling of wetlands [3]. Assessing anthropogenic alterations \n39 and mitigations needs up-to-date biological monitoring. Aquatic surveys are constrained by \n40 shoreline armoring, navigation restrictions, and submerged structures including cables, tunnels, \n41 bridge and pier supports, and rocky reefs. \n42 Environmental DNA may offer a way to advance the science and practice of \n43 biomonitoring [4]. The ease and nondestructive nature of collecting water for eDNA raises the \n44 prospect of monitoring biodiversity at fine scale in space and time. For instance, there are several \n45 metabarcoding primer sets targeting the mitochondrial 12S gene that amplify eDNA from most \n46 bony fish species [5-8]. One liter of estuary water typically contains sufficient eDNA to assay at \n47 least the more abundant taxa. Hindrances to wider adoption of metabarcoding for fish \n48 assessment include persistent uncertainty about how metabarcoding reads relate to eDNA \n49 abundance and how eDNA abundance relates to fish abundance [9,10]. Recent reports indicate \n50 progress addressing these essential concerns. First, there is increasing evidence that relative \n51 metabarcoding reads and eDNA levels are generally proportional to relative fish species \n52 abundance [11-15]. Second, this correlation can be improved by weighting metabarcoding reads \n53 to adjust for amplification bias, i.e., differences in PCR efficiency among species due to primer \n54 mismatch or other factors [16-19]. Alternatively, adding a known quantity of a synthetic DNA \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n4\n55 template as a standard to each PCR assay (“spike-in”) helps quantify absolute eDNA levels [20-\n56 22]. A related approach uses qPCR with metabarcoding primers to quantify total fish eDNA \n57 [23]. Quantification offers significant benefits. Absolute eDNA concentrations can be directly \n58 compared within and among studies. Spike-in quantification reveals limits to reproducible \n59 detection and adjusts for amplification of nontarget DNA [24]. Additional informative work \n60 regarding eDNA levels and fish abundance includes mesocosm testing of eDNA shedding and \n61 decay, effects of temperature and sunlight exposure on DNA degradation, modeling ocean \n62 dispersal including tidal effects, and accounting for allometric scaling [25-29]. Particularly in \n63 urban settings, further concerns arise regarding interference from wastewater DNA [30]. eDNA \n64 has inherent limitations as a biomonitoring tool, including absent information on life stage, age, \n65 health, weight, sex.\n66 In this report we applied spike-in metabarcoding with Riaz 12S gene primers to quantify \n67 vertebrate eDNA in New York City’s East River, a tidal strait in the lower Hudson River estuary \n68 [31]. The Hudson estuary has a storied history [32,33]. Environmental restoration efforts dating \n69 from the Clean Water Act of 1972 have begun to reverse centuries of neglect [34,35]. The East \n70 River study site challenges gear-based fish surveys because of the channel’s armored shoreline, \n71 irregular rocky bottom topography, and rapid tidal currents [36]. Wastewater permeates the \n72 estuary. Like many municipalities, New York City is served by a combined sewer system that \n73 conducts household sewage and stormwater from street runoff to underground reservoirs and \n74 then to treatment plants [37]. When conduits are overloaded, as frequently occurs in New York \n75 City after even modest rainfall, the effluent empties at combined sewer overflow (CSO) outfalls \n76 into waterways. CSO outfalls in New York City currently discharge about 18 billion gallons of \n77 untreated wastewater into the estuary annually. Although significant, this represents an 80% \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n5\n78 reduction in CSO discharges since 1985, thanks to more than $40 billion in infrastructure \n79 improvements [38]. We test two hypotheses: first, that estuary eDNA quantifies local marine fish \n80 populations without interference from wastewater DNA and second, that wastewater DNA \n81 usefully reports on other aspects of urban environment.\n82 Materials and Methods\n83 Replicate 1 L water samples were collected weekly at an East River shoreline site, used \n84 in prior studies, beginning May 2, 2024 through May 1, 2025 (Fig 1) [39]. As the shoreline is \n85 armored and elevated about 2 meters above water level, paired samples were obtained with \n86 separate throws of a bucket on a rope and transferred at site to laboratory bottles. Given tidal \n87 current up to 3 m/sec, estimated average water surface distance between samples was about 100 \n88 m (range 0 m – 400 m) [36]. Samples were brought to the laboratory within 15 min and each \n89 liter was filtered separately using a 4.5 cm, 0.45  M pore size nitrocellulose membrane \n90 (Millipore) and wall suction. Once complete, filters were folded to protect retained material and \n91 stored in 15 mL centrifuge tubes at -20C. Between uses, collection and filtration equipment \n92 were washed extensively with tap water and air dried. DNA was extracted from filters within \n93 three months using Qiagen DNeasy PowerSoil Pro Kit and eluted in 100 l of Buffer 6. DNA \n94 yield was measured with Qubit and samples were stored at -20C. Negative field controls \n95 consisted of 1 L samples of laboratory tap water filtered and extracted for DNA using the same \n96 equipment as for field samples. \n97\n98 Fig 1. Survey design. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n6\n99 Upper left, Collection location in New York City’s East River, a tidal channel between Long \n100 Island Sound and New York Harbor. Asterisk marks collection site. Right, map of lower Hudson \n101 River estuary and surrounds. Lower left, collection and analysis strategy. Map generated in \n102 Photoshop using USGS templates (https://apps.nationalmap.gov/viewer/). Photo credit Mark \n103 Stoeckle. \n104  \n105 PCR was done as previously described except that the spike-in standard was a 768 bp \n106 synthetic gene block (IDT), based on native ostrich 12S DNA amplicon standard in prior study \n107 (S1 Fig) [40]. Unlike the native ostrich 12S sequence, the synthetic gene block has M13 tails and \n108 three bases modified to match the MiFish-U forward primer [6]. Primary PCR assays were \n109 carried out with TaKaRa Titanium Taq hot start High-Yield EcoDry premix in 25 L volume \n110 with 200 M tailed Riaz primers, 250 copies gene block standard, and 5 L DNA or 5  L \n111 reagent-grade H2O, the latter as PCR negative control. Unlike other commonly used primer sets \n112 for fish metabarcoding, Riaz 12S primers are effective for most vertebrates, as 12S binding sites \n113 are conserved not only in bony fish but also in mammals and birds [5]. DNA samples were \n114 amplified in replicate and negative field and PCR controls were included in each amplification \n115 set. Primer sequences and thermal cycling conditions are in S1 Table. After amplification, 5  L \n116 of reaction mix were run on 2% agarose gel with SyberSafe to assess products and the remaining \n117 20 l were diluted 1:20 in Buffer EB and stored at -20C. Following procedure described in \n118 prior studies, 5 L of diluted primary PCR product were indexed with Nextera XT Index Kit v2 \n119 Set A and Cytiva PuReTaq Ready-to-Go PCR beads in 25  l volume. To assess products, 5 L \n120 of index reaction mix were run on a 2% agarose gel, and the remaining 20 L were pooled, \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n7\n121 cleaned with AMpure beads at 1:1, and resuspended in Buffer EB. Pooled libraries were \n122 sequenced at AZENTA using MiSeq, 2 x 150 bp, 10% PhiX, and 7.5 GB depth. Each library \n123 represented a single PCR run on a single DNA sample. The 192 field libraries plus 21 controls \n124 (11 field, 10 PCR) were analyzed in four sequencing runs together with unrelated libraries. To \n125 assess potential primer bias, a set of DNA samples (64 field samples plus negative controls) \n126 were amplified with MiFish-U-F/R2 primers and ostrich g-block spike-in as described above. \n127 MiFish-U-F/R2 primers have an extra 3’ base in reverse primer to reduce amplification of \n128 bacterial 16S DNA [40]. Primer sequences and thermal cycling parameters are in S1 Table. The \n129 protocol was otherwise the same as for Riaz primer amplifications. Pooled libraries were cleaned \n130 with AMpure and sequenced at AZENTA on a MiSeq, 2 x 250 bp with 10% PhiX. \n131 Bioinformatics used a DADA2 pipeline as previously described [24]. DADA2 output \n132 files were transferred to Excel. Amplicon sequence variants (ASVs) were filtered to exclude \n133 detections representing less than 0.1% of total reads for a given ASV. ASVs were identified by \n134 100% match to a local library of reference sequences representing local marine fish, local \n135 freshwater fish, nonlocal fish, non-fish wildlife, domesticated animal, and human (S2 Table). \n136 Unmatched sequences were manually submitted to GenBank using BLAST; additional matches \n137 were added to the reference library. For each library, reads per copy of ostrich standard were \n138 calculated. This proportion was then applied to convert reads to eDNA copies for all ASVs in \n139 that library (reads per ASV ÷ reads/copy standard = copies per ASV). As previously reported \n140 some local marine fish species shared Riaz segment 12S sequences, and some species were \n141 represented by more than one ASV (S2 Table). Results are expressed as eDNA copies per PCR \n142 assay or as per liter water sample, the latter obtained by multiplying copies per PCR by 20 to \n143 account for the proportion of DNA extract used for each PCR. For MiSeq fastq files, the pipeline \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n8\n144 was adjusted to accommodate the longer amplicon. Statistical tests were performed in GraphPad \n145 Prism 10.5.0. \n146 Results\n147 Vertebrate eDNA by category\n148 A total of 96 samples were obtained on 48 collecting days. eDNA was successfully \n149 extracted and amplified from all water samples and negative controls, generating 192 field \n150 sample and 21 negative control NGS libraries. ASVs in DADA2 output files were identified to \n151 species and assigned to vertebrate categories as described in Materials and Methods. Sequencing \n152 depth appeared sufficient to detect single copy eDNA. Average sequencing depth was 116,983 \n153 vertebrate reads per library (range 21,045 – 275,753); average reads per eDNA copy gene block \n154 standard were 88 (range 4 – 326) (S3 Table). Local marine fish eDNA was detected on all days \n155 (average copies/L, 12,507; range 1,161 – 58,738) but accounted for less than half of total \n156 vertebrate eDNA (Fig 2, S6 Table). The majority was human eDNA (average copies/L, 27,629; \n157 range, 1,295 – 279,273). Domesticated animal and other categories of vertebrate eDNA were \n158 commonly present. Negative control libraries yielded low levels of human (average copies/L, 24; \n159 range 0 – 173) and domesticated animal (average copies/L, 15; range 0 – 143) eDNA. (Fig 2, S4 \n160 Table). Other categories of vertebrate eDNA were not detected in negative controls. \n161\n162 Fig 2. Vertebrate eDNA abundance in field samples and negative controls. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n9\n163 Values represent daily averages (left) or individual negative controls (right). Note scale differs \n164 between graphs. Box indicates 25th-75th percentile, whiskers, 1.5x interquartile range, and \n165 outliers are shown as points; x marks average and line is median. Graph at left does not display \n166 outliers in human eDNA; plot otherwise represents values for complete dataset (S4 Table).\n167\n168 Local marine fish\n169 eDNA copy number reproducibility \n170 Species detection and eDNA copy number were largely reproducible in single PCR \n171 replicates, particularly for more abundant eDNA (S5 Table, S2 Fig). For species with more than \n172 10 copies/PCR, most (414/443; 94%) were detected in replicate library, whereas for those with \n173 fewer than 10 copies/PCR, only about half (124/278; 45%) were present. Excluding \n174 nonreplicated detections, the average absolute fold-difference in copies per species was modest \n175 (average, 2.0; standard deviation (SD),  2.5). As expected, there were larger differences \n176 between paired field collections (S5 Table, S2 Fig). About half of field sample detections above \n177 and below 10 copies/PCR were present in replicate library (225/443 (51%) and 125/278 (45%), \n178 respectively). Excluding non-detections, the average absolute fold-difference in copies per \n179 species was 3.1, SD,  8.2. Pooled PCR and field replicate datasets were reproducible over a \n180 wide concentration range (Fig 3). The average absolute fold-difference in copies per species was \n181 1.3 (SD  0.4) and 1.5 (SD  1.3), respectively (S6 Table). Amplification with MiFish-U-F/R2 \n182 primer set yielded copy numbers per species similar to that with Riaz primers (Fig 3; S7,S8 \n183 Tables; S3 Fig). One exception was cunner (Tautogolabrus adspersus), which amplified weakly \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n10\n184 with MiFish-U-F/R2 primers; this discrepancy was predictable based on primer mismatch (S4 \n185 Fig). \n186\n187 Fig 3. Reproducibility of pooled PCR and field replicates. \n188 Each point represents one local marine fish species in pooled sets of A) PCR or B) field \n189 replicates (96 PCRs/set). Sets are named as in Fig 1. Values are copies per pool. C) Fold-\n190 difference copies/species for pooled datasets generated with MiFish-U-F/R2 as compared to Riaz \n191 primers. Shading covers  4-fold difference. Asterisk indicates cunner (T. adspersus). Data \n192 sources in text.  \n193\n194 Species abundance distribution (SAD) \n195 Seventy-one local marine fish species were detected (S3 Table). Species yearly average \n196 eDNA abundance followed a hollow curve distribution ranging over four orders of magnitude \n197 (Fig 4). We propose an abundance scale based on order of magnitude differences in absolute \n198 eDNA concentration as shown in Fig 4. Abundant species were commonly detected and when \n199 detected, present in many copies (S5 Fig). Conversely, rare species were rarely detected and \n200 when detected, present in few copies. These two properties generated the hollow curve SAD. \n201 Abundant and common species (n = 16; 23% of species) were mostly reef-associated, consistent \n202 with the rocky nature of the East River site and accounted for the great majority (95%) of fish \n203 eDNA (Table 1).\n204\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n11\n205 Fig 4. Local marine fish SAD and proposed abundance categories. \n206 Each column represents one species. Species are rank ordered by average yearly abundance and \n207 shown in linear and log scale. \n208\n209 Table 1. Abundant, common local marine fish.\n210 Species are ranked by decreasing abundance.  \n211\neDNA copies/L\n% eDNA\nReef\nTautog Tautoga onitis 4298 34.4 Y\nCunner Tautogolabrus adspersus 1198 9.6 Y\n Menhaden, river herrings Brevoortia tyrannus, Alosa sp 1101 8.8\nSkilletfish Gobiesox strumosus 1068 8.5 Y\nFeather blenny Hypsoblennius hentz 531 4.2 Y\nStriped bass Morone saxatilis 506 4.0 Y\nBay anchovy Anchoa mitchilli 458 3.7\nOyster Toadfish Opsanus tau 439 3.5 Y\nSeaboard goby Gobiosoma ginsburgi 419 3.3 Y\nSpot Leiostomus xanthurus 401 3.2 Y\nBlack sea bass Centropristis striatus 362 2.9 Y\nAtlantic silverside Menidia menidia 299 2.4\nAmerican eel Anguilla rostrata 251 2.0\nScup Stenotomus chrysops 144 1.1 Y\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n12\nAtlantic herring Clupea harengus 132 1.1\nTomcod Microgadus tomcod 108 0.9\nBluefish Pomatomus saltatrix 104 0.8\nSUM 11819 94.5\n212\n213\n214 Phenology\n215 There was 10-fold seasonal variation in total marine fish eDNA which roughly paralleled \n216 seasonal variation in New York Harbor water temperature (Fig 5).  There was a similar seasonal \n217 pattern in daily species richness, such that no species were abundant in winter and few were \n218 common. Daily species richness for uncommon and rare species did not show a clear seasonal \n219 trend. Individual species differed in seasonality (Fig 6). \n220\n221 Fig 5. Local marine fish eDNA seasonal abundance and diversity. \n222 Each column represents one collection day. Collection months, colored by quarter, are indicated; \n223 overlay shows New York Harbor water temperature. Left, total copies/L local marine fish. Right, \n224 number of species, colorized by abundance rank on collection day as in Fig 4 (source data S3 \n225 Table). \n226\n227 Fig 6. Seasonal eDNA abundance individual species. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n13\n228 Selected species with differing phenology are shown. Each column represents one collection \n229 day; scale is copies/L. Scale differs between graphs. Color bar indicates months as in Fig 5. \n230 Scientific names, source data in S3 Table.\n231\n232 Other vertebrate eDNA\n233 Human, domesticated animal, non-fish wildlife, and nonlocal fish eDNA was commonly \n234 detected in estuary samples (Fig 2). Unlike local marine fish eDNA, daily abundances of other \n235 vertebrate eDNA categories were directly correlated to that for human eDNA, consistent with a \n236 shared wastewater source (Fig 7). Estuary eDNA correctly identified the commonest urban \n237 mammals, land birds, and household pets (Table 2) [41,42]. Dietary animal eDNA proportions \n238 closely tracked proportions in national statistics on meat and fish consumption (Fig 8) [43-46].\n239\n240 Fig 7. Other vertebrate eDNA abundance by category in relation to human eDNA \n241 abundance. \n242 Each point represents one collection day. Scale is log10 copies/L. \n243\n244 Table 2. Top species other vertebrate categories.  \n245\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n14\nCopies/L\nDetections\nNon-fish wildlife (62) Mammals (22) Rat 684 40\nGray squirrel 23 20\nRaccoon 19 14\nBeaver 13 14\nHouse mouse 12 14\nWhite-tailed deer 11 10\nDeer mouse 11 10\nLand birds (19) Rock pigeon 254 36\nStarling 19 11\nHouse sparrow 11 8\nWaterbirds (19) Canada goose 219 33\nRing-billed gull 89 26\nHerring gull, other Larus sp 74 30\nMallard, other Anas sp 59 33\nDC cormorant 42 31\nGreater scaup, other Athya sp 11 8\nDomesticated animal (16) Dietary (11) Chicken 1601 47\nCow 867 46\nPig 716 43\nTurkey 57 29\nSheep 56 26\nGoat 26 14\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n15\nPets (5) Dog 252 40\nCat 104 22\nNonlocal fish (55) Atlantic salmon 80 30\nEuropean sea bass 43 16\nTilapia 24 12\nRainbow trout, other Oncorhynchus sp 20 13\nRed snapper, other Lutjanus sp 19 7\nHuman 27629 192\n246\n247 Average copies/L and days detected (total collection days n = 48) are shown. Taxa are ranked by \n248 decreasing abundance within each category; rare species (<10 copies/L) not shown. Total \n249 detected in parentheses. Scientific names, complete dataset in S3 Table. \n250\n251 Fig 8. Proportions dietary animal eDNA compared with proportions US consumption per \n252 capita. \n253 Data sources in text.\n254\n255 Discussion\n256 In this investigation we employed spike-in metabarcoding to quantify vertebrate eDNA \n257 weekly for one year in an urban estuary. There were two main findings. First, results supported \n258 hypothesis that eDNA indexes local marine fish abundance. eDNA followed expected fish \n259 survey characteristics including a classic hollow-curve species abundance distribution, reef \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n16\n260 specialists as top species, summer peak in total fish abundance, and species-specific seasonal \n261 patterns consistent with life histories [47-51]. Human and other wastewater eDNA was plentiful \n262 but total vertebrate eDNA levels were below the threshold expected to suppress fish assessment. \n263 Second, wastewater eDNA tracked terrestrial wildlife abundance and human consumption of \n264 meat and fish.\n265 We propose a fish abundance scale based on order of magnitude differences in absolute \n266 eDNA concentration (Fig 2). An order of magnitude scale has been applied to rank numerical \n267 abundance of fish species [52]. Standardized numerical categories of eDNA abundance could \n268 help communicate survey findings to general as well as scientific audiences. eDNA rarity was \n269 the main limit to eDNA detection and quantification, a recognized constraint in fish eDNA \n270 surveys [53-55]. The protocol employed a PCR input of 1/20th of the DNA obtained from 1 liter \n271 of water, a similar proportion as that in other studies. Fish eDNA copies per PCR aliquot were \n272 surprisingly low (average, 625; range, 21 – 3271). This was typically sufficient to detect a dozen \n273 or so species (average, 15; range, 4 – 26).  Across the one-year survey, most of 71 local marine \n274 fish species were present in fewer than 10% of PCRs (S3 Table). At the other end of the \n275 distribution, skilletfish (G. strumosus) and feather blenny (H. hentz) were newly abundant (Table \n276 1). These taxa were rare in an eDNA survey at this site in 2016 and increased in 2022 (Fig 9) \n277 [39,56]. Both species were rare or absent in regional surveys up to 2020 [52,57-59]. Hudson \n278 River Park Fish Survey (HRPFS) corroborates recent increases [60]. Skilletfish and feather \n279 blenny were first recorded in HRPFS traps in 2020 and subsequently commonly collected. Both \n280 species favor oyster reef habitat; the current plenitude might be consequent to restoration of \n281 oyster beds in New York Harbor that began in 2015 [61]. \n282\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n17\nFig 9. Newly abundant species. \nData are from this study and previously reported East River eDNA surveys (see text for \nreferences).\n283\n284 Wastewater DNA offered insights. Daily levels of non-fish vertebrate eDNA correlated \n285 with daily levels of human DNA, consistent with a shared wastewater source. The highest levels \n286 of human and other vertebrate eDNA were obtained after significant rainfall, evidence that peak \n287 concentrations resulted from CSO discharge of untreated waste (S6 Fig). Whether processed \n288 wastewater contributes human and other vertebrate eDNA to the Hudson estuary is unknown. It \n289 is recently reported that dietary fish eDNA persists even after quaternary treatment of sewage \n290 [62]. Future work could examine eDNA content of untreated sewage and street runoff at CSO \n291 outfalls, at different points in the processing system, and elsewhere in the estuary. Human and \n292 domesticated animal DNA may serve as markers of human environmental impact [63]. \n293 Wastewater is often considered a nuisance in eDNA surveys because of possible suppression of \n294 fish reads by human DNA and misleading detection of dietary species [30,64]. Some studies use \n295 human blocking primers, which have unknown effects on amplification of target taxa. Present \n296 results suggest this precaution may be unnecessary, the more so if primers are selective for fish \n297 vs. other vertebrates. In this report the apparent copy number of human DNA as assayed with \n298 MiFish-U-F/R2 primers, which have multiple binding site mismatches in mammals, was about \n299 1/1000th of that obtained with Riaz primers (S8 Table). \n300 Rats are the most abundant wild mammal in New York City [65,66]. Monitoring rat \n301 eDNA in the estuary may help assess pest control programs. Standardizing against other \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n18\n302 terrestrial wildlife eDNA such as pigeon might compensate for the variable content of street \n303 runoff in estuary samples, or street runoff could be directly sampled. Routine estuary eDNA \n304 testing could greatly inform urban wildlife management. More than 60 non-fish wildlife species \n305 were identified including fauna rarely seen (S3 Table). Nearly all were taxa known to be resident \n306 within city limits, although transport from distant sites or local zoos cannot be excluded [41,42]. \n307 Proportions of livestock and dietary fish eDNA in wastewater closely approximated proportions \n308 in national consumption statistics. To our knowledge this is the first demonstration that \n309 wastewater eDNA reports on human diet, a finding of potential value to public health and \n310 commercial interests. The proportions of goat and sheep in East River eDNA were greater than \n311 in national data, which might reflect increased consumption by ethnic populations in New York \n312 City. This hypothesis could be explored with local sales data. \n313 Limitations\n314 This report has several limitations. A concurrent gear-based survey is not available for \n315 benchmarking eDNA. Assessment of eDNA performance for fish censusing rests on features \n316 outlined above. Sampling was conducted at a single location, so findings incompletely depict \n317 lower Hudson River estuary fish populations. Nonetheless, fish eDNA diversity broadly \n318 overlapped with traditional surveys conducted in the estuary over the past 35 years [52,57-59]. \n319 Potential tidal effects were not directly assessed. Given semidiurnal tides and weekly water \n320 sample collection, consecutive samples were drawn on approximately opposite tides. Inspection \n321 did not show consistent week-to-week variation in eDNA profiles, and exploratory plots of \n322 eDNA copies/species vs. tide timetable were similarly unrevealing. The apparent absence of \n323 differences may reflect limited dispersion, as tidal excursion in the East River is less than the \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n19\n324 length of the channel [67]. Many fish species spawn in the estuary [35]. Possible contributions of \n325 spawning, larval, or juvenile fish to eDNA levels are unknown. Water temperature might alter \n326 eDNA shedding or decay rates [26]. However, such across-the-board factors insufficiently \n327 account for species-specific phenologies obtained with eDNA. eDNA persistence beyond weekly \n328 sampling intervals might blunt reporting on current fish abundance. Individual species \n329 demonstrated strong seasonal patterns consistent with known phenology, consistent with \n330 inference that the assay indexes current or recent fish density (Fig 6). In addition, peak values of \n331 human and domesticated animal eDNA associated with recent rainfall did not persist into the \n332 following week (S6 Fig). Assignment of species to categories carries uncertainties. Some local \n333 marine species are also consumed by humans—striped bass (M. saxatilis), for example. eRNA \n334 assays may enable distinguishing nucleic acid signals due to resident fish from those introduced \n335 by waste [68]. Nonlocal fish eDNA attributed to wastewater may have originated from \n336 extralimital strays. \n337 Reproducibility and accuracy are desirable attributes in surveying marine biodiversity. \n338 These considerations apply to both questions raised in the Introduction—whether metabarcoding \n339 reports on eDNA levels and whether eDNA levels report on fish abundance. Replication \n340 experiments demonstrated good reproducibility in measuring eDNA levels, particularly for more \n341 abundant eDNAs. Accuracy might be distorted by differences among species in amplification \n342 efficiency. Quantitatively similar results for most species with MiFish-U-F/R2 as compared with \n343 Riaz primers suggest this is not a major factor for bony fish (Fig 3, S3 Fig). Cunner (T. \n344 adspersus) was an exception. The apparent concentration of cunner eDNA was about 70-fold \n345 lower with MiFish-U-F/R2 than with Riaz primers, a deficit predicted by binding site mismatch \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n20\n346 (S4 Fig). Dietary animal eDNA added evidence that the Riaz assay accurately reported relative \n347 eDNA levels. \n348 There are limits to knowledge about abundance of marine biodiversity [69]. Censusing is \n349 inherently imprecise for fish and other nekton [70]. Given that fish SADs typically extend over \n350 four or five orders of magnitude, a reproducible protocol that indexes eDNA or fish abundance \n351 within, say, half an order of magnitude (about threefold) of the reference value would have wide \n352 use. There is no certain gold standard in benchmarking eDNA for marine fish assessment—all \n353 methods have catchability biases [71,72]. Estuary eDNA detected 22 freshwater fish species, all \n354 uncommon or rare (S3 Table); these findings may be of interest for future work. Waterbird \n355 eDNA was grouped with that of other non-fish wildlife but probably originated from birds in the \n356 estuary rather than from wastewater; this could be tested directly. The survey required about \n357 25% effort and direct costs of about $15,000 (S9 Table). About half of effort was devoted to \n358 bioinformatics; this might be reduced by automating some procedures performed in Excel. \n359 Conclusion\n360 Vertebrate eDNA metabarcoding with spike-in quantification offers a practical approach \n361 to biomonitoring in urban estuaries. This methodology holds promise as an aid to estuary fish \n362 and wildlife management and opens a window into human diet. Regular reporting of such \n363 findings would likely be of interest to both government and non-government entities [73]. \n364 Absolute quantification of eDNA levels could yield heretofore unprecedented ability to map fish \n365 abundance across diverse sites and habitats. \n366\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n21\n367 Acknowledgments\n368 We thank Jeanne Garbarino, Jen Bohn, and Jessi Hersh for generous assistance and \n369 sharing laboratory supplies and equipment, and David Thaler for helpful comments on \n370 manuscript.  \n371\n372 References\n373\n374 1. Sanderson EW. Mannahatta: A natural history of New York City. New York: Abrams; 2009. \n375 2. Pinto PJ, Kondolf GM. Evolution of two urbanized estuaries: environmental change, legal \n376 framework, and implications for sea-level rise vulnerability. Water. 2016;8: 535. \n377 3. Waltham J, McCann J, Power T, Moore M, Buelow C. Patterns of fish use in urban estuaries: \n378 engineering maintenance schedules to protect broader seascape habitat. Estuar Coast Shelf Sci. \n379 2020;238: 106729. \n380 4. Taberlet P, Bonin A, Zinger L, Coissac E. Environmental DNA for biodiversity research and \n381 monitoring. New York: Oxford University Press; 2018.\n382 5. Riaz T, Shehzad W, Viari A, Pompanon F, Taberlet P, Coissac E. ecoPrimers: inference of \n383 new DNA barcode markers from whole genome sequence analysis. Nucleic Acids Res. 2011;39: \n384 e145.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n22\n385 6. Miya M, Sato Y, Fukunaga T, Sado T, Poulsen JY, Sato K, et al. MiFish: a set of universal \n386 PCR primers for metabarcoding environmental DNA from fishes: detection of more than 230 \n387 subtropical marine species. R Soc Open Sci. 2015;2: 150088.\n388 7. Valentini A, Taberlet P, Miaud C, Civade R, Herder J, Thomsen PF, et al. Next-generation \n389 monitoring of aquatic biodiversity using environmental DNA metabarcoding. Mol Ecol. \n390 2016:25: 929-942.\n391 8. Zhang S, Zhao J, Yao M. A comprehensive and comparative evaluation of primers for \n392 metabarcoding eDNA from fish. Methods Ecol Evol. 2020;11: 9-25.\n393 9. Hansen B, Bekkevold D, Clausen LW, Nielsen EE. The sceptical optimist: challenges and \n394 perspectives for the application of environmental DNA in marine fisheries. Fish Fish. 2018;19: \n395 751-768.\n396 10. Darling JA, Jerde CL, Sepulveda AJ. What do you mean by false positive? Environ DNA. \n397 2021;3: 879-883.\n398 11. Salter I, Joensen M, Kristiansen R, Steingrund P, Vestergaard P. Environmental DNA \n399 concentrations are correlated with regional biomass of Atlantic cod in oceanic waters. Commun \n400 Biol. 2019;2: 461. \n401 12. Stoeckle MY, Adolf J, Charlop-Powers Z, Dunton KJ, Hinks G, VanMorter SM. Trawl and \n402 eDNA assessment of marine fish diversity seasonality and relative abundance in coastal New \n403 Jersey USA. ICES J Mar Sci. 2021;78: 293–304.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n23\n404 13. Shelton AO, Ramón-Laca A, Wells A, Clemons J, Chu D, Feist BE, et al. Environmental \n405 DNA provides quantitative estimates of Pacific hake abundance and distribution in the open \n406 ocean. Proc R Soc Lond B Biol Sci. 2022;289: 20212613.\n407 14. Jiang P, Zhang S, Xu S, Xiong P, Cao Y, Chen Z, et al. Comparison of environmental DNA \n408 metabarcoding and bottom trawling for detecting seasonal fish communities and habitat \n409 preference in a highly disturbed estuary. Ecol Indic. 2023;146: 109754.\n410 15. Kasmi Y, Blancke T, Eschbach E, Möckel B, Casas L, Bernreuther M, et al. Atlantic cod \n411 (Gadus morhua) assessment approaches in the North and Baltic Sea: A comparison of \n412 environmental DNA analysis versus bottom trawl sampling. Front Mar Sci. 2023;10: 58354.\n413 16. Shelton AO, Gold ZJ, Jensen AJ, D’Agnese E, Allan EA, Van Cise A, et al. Toward \n414 quantitative metabarcoding. Ecol. 2022;104: e3906. \n415 17. Ledger KJ, Hicks MBR, Hurst TP, Larson W, Baetscher DS. Validation of environmental \n416 DNA for estimating proportional and absolute biomass. Environ DNA. 2024;6: e70030.\n417 18. Guri G, Shelton AO, Kelly RP, Yoccoz N, Johansen T, Præbel K, et al. Predicting trawl \n418 catches using environmental DNA. ICES J Mar Sci. 2024;81: 1536-1548.\n419 19. Shaffer MR, Allan EA, Van Cise AM, Parsons KM, Shelton AO, Kelly RP. Observation bias \n420 in metabarcoding. Mol Ecol Resour. 2025; e14119.\n421 20. Ushio M, Murakami H, Masuda R, Sado T, Miya M, Sakurai S, et al. Quantitative \n422 monitoring of multispecies fish environmental DNA using high-throughput sequencing. \n423 Metabarcoding Metagenom. 2018;2: 1-15.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n24\n424 21. Sato M, Inoue N, Nambu R, Furuichi N, Imaizumi T, Ushio M. Quantitative assessment of \n425 multiple fish species around artificial reefs combining environmental DNA metabarcoding and \n426 acoustic survey. Sci Rep. 2021;11: 19477. \n427 22. Tsuji S, Inui R, Nakao R, Miyazono S Saito M, Kono T, et al. Quantitative environmental \n428 DNA metabarcoding shows high potential as a novel approach to quantitatively assess fish \n429 community. Sci Rep. 2022;12: 21524. \n430 23. Pont D, Meulenbroek P, Bammer V, Dejean T, Erõs T, Jean P, et al. Quantitative monitoring \n431 of diverse fish communities on a large scale combining eDNA metabarcoding and qPCR. Mol \n432 Ecol Resour. 2023;23: 396-409.\n433 24. Stoeckle MY, Ausubel JH, Coogan M. 12S gene metabarcoding with DNA standard \n434 quantifies marine bony fish environmental DNA identifies threshold for reproducible detection \n435 and overcomes distortion due to amplification of non-fish DNA. Environ DNA. 2022;6: e376. \n436 25. Andruszkiewicz EA, Koseff JR, Fringer OB, Ouelette NT, Lowe AB, Edwards CA, et al. \n437 Modeling environmental DNA transport in the coastal ocean using Lagrangian particle tracking. \n438 Front Mar Sci. 2019;6: 477. \n439 26. Allan EA, Zhang WG, Lavery AC, Govindarajan AF. Environmental DNA shedding and \n440 decay rates from diverse animal forms and thermal regimes. Environ DNA. 2020;3: 492–514. \n441 27. Yates MC, Glaser DM, Post JR, Cristescu ME, Fraser DJ, Derry AM. The relationship \n442 between eDNA particle concentration and organism abundance in nature is strengthened by \n443 allometric scaling. Mol Ecol. 2021;30: 3068-3082.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n25\n444 28. Mauvisseau Q, Harper LR, Sander M, Hanner RH, Kleyer H, Deiner K. The multiple states \n445 of environmental DNA and what is known about their persistence in aquatic environments. \n446 Environ Sci Technol. 2022;56: 5322-5333.\n447 29. Baetscher DS, Pochardt MR, Barry PD, Larson WA. Tide impacts the dispersion of eDNA \n448 from nearshore net pens in a dynamic high-latitude marine environment. Environ DNA. 2024;6: \n449 e533.\n450 30. Zhan A. Overlooked eDNA contamination in human-influenced ecosystems: a call to \n451 manage large-scale false-positives in biodiversity assessments. Water Biol Secur. 2025; \n452 1003754.\n453 31. Levinton JS, Waldman JR, eds. The Hudson River Estuary. New York: Cambridge \n454 University Press; 2006.\n455 32. Shorto R. The island at the center of the world. 2025. New York: Vintage; 2025.\n456 33. Kurlansky M. The big oyster: history on the half shell. New York: Ballantine; 2006.\n457 34. Waldman J. Heartbeats in the muck: the history, sea life, and environment of New York \n458 Harbor, Revised edition. New York: Empire State Editions; 2012. \n459 35. The Hudson River Estuary Management Program and NY-NJ Harbor & Estuary Program. \n460 The state of the estuary 2025. New York. November 2024. Accessible at \n461 https://www.hudsonriver.org/ccmp/soe.\n462 36. Bowman MJ. The tides of the East River, New York. J Geophys Res Oceans. 1976;81: 1609-\n463 1616.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n26\n464 37. New York State Department of Environmental Conservation. Combined sewer overflow. \n465 Accessible at https://dec.ny.gov/environmental-protection/water/water-quality/combined-sewer-\n466 overflow.\n467 38. New York City Department of Environmental Protection. Combined sewer overflow long \n468 term control plan for citywide/open waters. Executive summary, September 2020.  Accessible at \n469 https://www.nyc.gov/assets/dep/downloads/pdf/water/nyc-waterways/citywide-east-river-open-\n470 water/citywide-open-waters-ltcp-executive-summary.pdf.\n471 39. Stoeckle MY, Soboleva L, Charlop-Powers Z. Aquatic environmental DNA detects seasonal \n472 fish abundance and habitat preference in an urban estuary. PLoS One. 2017;12: e0175186.\n473 40. Stoeckle MY, Adolf J, Ausubel JH, Charlop-Powers Z, Dunton KJ, Hinks G. Current \n474 laboratory protocols for detecting fish species with environmental DNA optimize sensitivity and \n475 reproducibility especially for more abundant populations. ICES J Mar Sci. 2022;79: 403–412.\n476 41. Mandelbaum R, Beck C. Wild NYC: experience the amazing nature in and around New \n477 York City (Wild Series). Portland; Timber Press: 2025. \n478 42. New York City Department of Parks & Recreation. Wildlife in New York City. Accessible at \n479 https://www.nycgovparks.org/learn/wildlife-in-new-york-city.\n480 43. National Chicken Council. Per capita consumption of poultry and livestock, 1965 to forecast \n481 2026. Accessible at https://www.nationalchickencouncil.org/about-the-industry/statistics/per-\n482 capita-consumption-of-poultry-and-livestock-1965-to-estimated-2012-in-pounds/.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n27\n483 44. U.S. Department of Agriculture Economic Research Service. Sheep, lamb & mutton—sector \n484 at a glance. Accessible at https://www.ers.usda.gov/topics/animal-products/sheep-lamb-\n485 mutton/sector-at-a-glance.\n486 45. National Marine Fisheries Service (2024). Fisheries of the United States, 2022. U.S. \n487 Department of Commerce, NOAA current fisheries statistics, No. 2022. Accessible at \n488 https://www.fisheries.noaa.gov/national/sustainable-fisheries/fisheries-united-states.\n489 46. National Fisheries Institute. Top 10 list for seafood consumption. Accessible at \n490 https://aboutseafood.com/news/nfi-annual-top-10-list-looks-at-seafoods-most-consumed-species-\n491 in-2022/.\n492 47. McGill BJ, Etienne RS, Gray JS, Alonso D, Anderson MJ, Benecha HK, et al. Species \n493 abundance distributions: moving beyond prediction theories to integration within an ecological \n494 framework. Ecol Lett. 2007;10: 995-1015.\n495 48. Able KW, Fahay MP. The first year in the life of estuarine fishes in the Middle Atlantic \n496 Bight. New Brunswick: Rutgers University Press; 1998.\n497 49. Collette BB, Klein-MacPhee G, eds. Fishes of the Gulf of Maine, Third Edition. Washington, \n498 D.C.: Smithsonian Institution Press; 2002.\n499 50. Froese R, Pauly D, eds. FishBase, World Wide Web electronic publication. Accessible at \n500 www.fishbase.org.\n501 51. Murdy EO, Birdsong RS, Musick JA. Fishes of Chesapeake Bay. Washington, D.C.: \n502 Smithsonian Institution Press; 1997.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n28\n503 52. Anadón JD, Piñeiro O, Ruhi A, Hornstein J, Waldman JR. Decoupled shifts of dominant and \n504 rarer fish species as a response to warming and extreme events in a large estuary. Ecosphere. \n505 2024;15: e4876.\n506 53. Govindarajan AF, McCartin L, Adams A, Allan E, Belani A, Francolini R, et al. Improved \n507 biodiversity detection using a large-volume environmental DNA sampler with in situ filtration \n508 and implications for marine eDNA sampling strategies. Deep Sea Res I. 2022;189: 103871. \n509 54. McClenaghan B, Fahner N, Cote D, Chawarski J, McCarthy A, Rajabi H, Singer G, et al. \n510 Harnessing the power of eDNA metabarcoding for the detection of deep-sea fishes. PLoS One. \n511 2020;15: e0236540.\n512 55. Shirazi S, Meyer RS, Shapiro B. Revisiting the effect of PCR replication and sequencing \n513 depth on biodiversity metrics in environmental DNA metabarcoding. Ecol Evol. 2021;11: \n514 15766-15779.\n515 56. Stoeckle MY, Ausubel JH, Hinks G, VanMorter SM. A potential tool for marine \n516 biogeography: eDNA dominant species differ among coastal habitats and by season concordant \n517 with gear-based assessments. PLoS One. 2024;19:e0313170.\n518 57. US Army Corps of Engineers. Demersal fish assemblages of New York/New Jersey Harbor \n519 and near shore fish communities of New York Bight, October 2015. Accessible at \n520 https://www.nan.usace.army.mil/Portals/37/docs/harbor/Biological%20and%20Physical%20Mo\n521 nitoring/ABS%20Multiple%20Summary%20Reports/NYD_ABS_Demersal_Fish_Assemblages\n522 _FINAL_14October2015_V2GraphicUpdate.pdf.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n29\n523 58. New York City Economic Development Corporation Financial District and Seaport Climate \n524 Resilience Plan: Biological and habitat sampling report, 2020-2021. Accessible at \n525 https://edc.nyc/sites/default/files/2024-03/FiDi-Master-Plan-East-River-Sampling-Testing-\n526 Year1-Findings.pdf.\n527 59. Park PJ, Girgenti CD, Del Bello IG, Tobitsch CM, Gorsen DM, Stanner KC, et al. New York \n528 City East River fish species inventory and emergence of a unique fish community science \n529 network. Urban Nat. 2020;38: 1-27. \n530 60. Hudson River Park Trust. Hudson River Park fish survey report 2024. Accessible at \n531 https://hudsonriverpark.org/app/uploads/2025/01/2024_Fish-Survey-Report_FINAL.pdf.\n532 61. Billion Oyster Project, 2025. Restoring oyster reefs to New York Harbor through public \n533 education initiatives. Accessible at https://www.billionoysterproject.org.\n534 62. Inoue Y, Miyata K, Yamane M, Honda H. Environmental nucleic acid pollution: \n535 characterization of wastewater generating false positives in molecular ecological surveys. ACS \n536 EST Water. 2023;3: 756-764. \n537 63. Thaler DS, Ausubel JH, Stoeckle MY. Human and domesticated animal environmental DNA \n538 as bioassays of the Anthropocene. Innovation. 2022;4: 100356. \n539 64. Xiong W, MacIsaac HJ, Zhan A. An overlooked source of false-positives in eDNA-based \n540 biodiversity assessment and management. J Environ Manag. 2024;358: 120949.\n541 65. Auerbach J. Does New York City really have as many rats as people? Signif. 2014;11: 22-\n542 27.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n30\n543 66. Wong T. There are 3 million rats in NYC, a 50% increase since 2010. 2025. Accessible at \n544 https://mandmpestcontrol.com/pests/rats/3-million-rats-in-nyc/.\n545 67. Blumberg AF, Pritchard DW. Estimates of the transport through the East River, New York. J \n546 Geophys Res Oceans. 1997;102: 5685-5703.\n547 68. Miyata K, Inoue Y, Amano Y, Nishioka T, Yamane M, Kawaguchi T, et al.  Fish \n548 environmental RNA enables precise ecological surveys with high positive predictivity. Environ \n549 Indic. 2021;128: 107796. \n550 69. Ausubel JH. On the limits to knowledge of future marine biodiversity. Elec J Sustain Dev. \n551 2008;1: 19-23.\n552 70. Privitera-Johnson K, Punt AE. A review of approaches to quantifying uncertainty in fisheries \n553 stock assessments. Fish Res. 2020;226: 105503.\n554 71. Arregúin-Sánchez F. Catchability: a key parameter for fish stock assessment. Rev Fish Biol \n555 Fish. 1996;6: 221-242.  \n556 72. Fraser HM, Greenstreet SPR, Piet GJ. Taking account of catchability in groundfish survey \n557 trawls: implications for estimating demersal fish biomass. ICES J Mar Sci. 2007;9: 1800-1819. \n558 73. Kelly RP, Lodge DM, Lee KN, Theroux S, Sepuveda AJ, Scholin CA, et al. Toward a \n559 national eDNA strategy for the United States. Environ DNA. 2024;6: e432.\n560\n561 Supporting Information\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n31\n562 S1 Table. PCR primers, protocols.\n563 S2 Table. Reference sequences. \n564 S3 Table. eDNA copies per ASV by PCR, day.\n565 S4 Table. eDNA copies per day by category for field samples, negative controls.\n566 S5 Table. Reproducibility individual PCR and field replicates.\n567 S6 Table. Reproducibility pooled PCR and field replicates.\n568 S7 Table. MiFish-U-F/R2 copies/ASV.\n569 S8 Table. MiFish-U-F/R2 vs Riaz copies/ASV.\n570 S9 Table. Effort, costs.\n571 S1 Fig. Gene block spike-in standard.\n572 S2 Fig. Reproducibility of individual PCR and field replicates.\n573 S3 Fig. MiFish-U-F/R2 vs Riaz pooled copies/L. \n574 S4 Fig. Primer binding sites for local marine fish species detected in this study. \n575 S5 Fig. Frequency of detection, copies per detection vs eDNA overall abundance.\n576 S6 Fig. Human eDNA abundance and recent rainfall.\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted September 8, 2025. ; https://doi.org/10.1101/2025.09.04.674353doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}