|
Article
<br /> Methods Length-weight fitting parameters were obtained from a comprehen-
<br /> sive assessment of Hawai'i specific parameters56 and FishBase65.Fish
<br /> Study site species were excluded from fish biomass calculations according to life
<br /> Hawai'i Island(19.550 N,155.660 W)is the southeastern most island of history characteristics that are not well captured with visual surveys,
<br /> the Hawaiian Archipelago,located in the northern central Pacific(Fig.1). including cryptic benthic species,nocturnal species,pelagic schooling
<br /> The western section has roughly 200 km of coastline predominantly species and manta rays.
<br /> oriented north to south.The coastline contains the longest contiguous
<br /> reef ecosystem in the main Hawaiian Islands55 and large gradients in Human population.We quantified human population density using
<br /> human population,local land-sea impacts and environmental factors NASA Gridded Population of theWorld v.4(ref.66).The dataset is avail-
<br /> that arecomparable to reef ecosystems globally(Extended Data Fig.1). ableatl km resolution at5-year intervals.Linear interpolation was used
<br /> The region represents an ideal studylocation for resolvingthe land-sea to fill in the missing years and produce annual time steps of human
<br /> human impacts driving reef ecosystem change and coral trajectories population within 15 km of each 100 mgrid cell across our study region
<br /> following acute climate-driven disturbance. (Supplementary Fig.12).
<br /> Reef surveys Wastewater pollution.We calculated wastewater effluent(I ha-'yr')
<br /> Full details related to sampling design,site selection and survey and nitrogen input(kg ha-'yr')from onsite sewage disposal systems
<br /> frequency for benthic and reef-fish data collection across our study (for example,cesspools and septic tanks)and injection wells(collec-
<br /> region are in the Supplementary Information.In brief,underwater tivelyOSDS)in coastal waters at100 m resolution.OnlyOSDS located
<br /> visual surveys of benthic assemblages were collated from three moni- within a modelled one-year groundwater travel time of the coast were
<br /> toring programmes for the following years(number of reefs surveyed included in the analysis and nutrients from OSDS were assumed to
<br /> are in parentheses):2003(23),2007(23),2011(23),2014(40),2015 flow to the nearest point on the shoreline.Wastewater effluent and
<br /> (40),2016(80),2017(80),2018(15)and 2019(55).All benthic surveys nutrient input were estimated on thebasis of ref.67and discharge rates
<br /> used permanently marked pins to ensure the same area of reef was and nutrient loading according to ref.68.A Gaussian decay function
<br /> surveyed over time.High-resolution photographs were collected by was used to estimate dispersal offshore,approaching zero at 2 km
<br /> using photoquadrats at 1 m intervals along 25 m belt-transects(n=26 (Supplementary Figs.13-15).This samedispersal function wasalsoused
<br /> photographs per transect).Between 30 and 50 random points were for nutrient input,urban runoff,sediment input and rainfall,which are
<br /> overlaid on each photograph and the benthic component under each each described below.
<br /> point was identified to the lowest possible taxonomic level.Percent-
<br /> age cover of the major functional groups at each reef were used in this Nutrient input.We calculated nutrient input(kg ha-'yr')at100m
<br /> analysis,namely hard coral and crustose coralline algae.Surveys of resolution as the combination of total nitrogen from OSDS(Waste-
<br /> reef-fish assemblages were performed along the same permanently water pollution section above)and golf courses.The total golf course
<br /> marked 25 m transects concurrently with benthic surveys.However, area per watershed was derived from NOAA Coastal Change Analysis
<br /> reef-fish surveys were performed more frequently(one to six times Program(CCAP)land-useand land-coverdata and Landsat cloud-free
<br /> per year from 2003 to 2019)than benthic surveys,depending on the composite images created with Google Earth Engine.The golf course
<br /> reef location and monitoring programme performing the surveys.In area was multiplied byan annual nitrogen application rateof 585 kg ha-'
<br /> all surveys,fishes were identified to species,sized and enumerated. (refs.69,70)and then by leaching rate of32%7-73to estimate nitrogen
<br /> To account for differences among programmes in how researchers that either runs off or reaches the groundwater.We also imposed a
<br /> surveyed reef fish,counts were calibrated using species and method reduction in nitrogen that reached the ocean on the basis of distance
<br /> specific adjustments56. inland and used subwatershed catchment data'to estimate nutrient
<br /> transport from golf courses to the coastline(Supplementary Figs.16-18).
<br /> Local land-sea human impacts and environmental factors
<br /> Fish biomass.The biomass of fishes at a given reef was measured as Urban runoff.We quantified the total area of impervious surfaces
<br /> total fishbiomass,herbivore fish biomass and the biomass of browsers, (that is,paved roads,parking lots,sidewalks and roofs)within 10 km
<br /> grazers and scrapers56.Total fish biomass is an indicator of the overall of the coastline at 100 m resolution for each year from 2000 to 2017
<br /> stateofthefish assemblage 57 and is reduced inareasthat have increased (Supplementary Figs.19 and 20).Data wereextracted from NOAACCAP
<br /> fishing pressures8,s9.InHawai'i,non-commercial nearshore fisheries land-use land-coverdata from1992,2001,2005and 2010.We also digi-
<br /> dominate,with people fishing for recreational,subsistenceand cultural tized 2017 impervious surface cover from a single cloud-free Landsat
<br /> purposes60,61 However,the dominant harvesting modesand magnitude 8 image(courtesyof the United States Geological Survey,USGS)(15 m
<br /> of fishing activities are largely unknown at spatial or temporal scales resolution pan-sharpened).Years in between data availability were
<br /> relevant to this stud y62.As such,we include total fish biomass in part filled in by linear interpolation.
<br /> to represent fishing effort on reefs but recognize its shortcomings in
<br /> capturing reef-and species-specific differences in fishing pressure Rainfall.Wequantified annual rainfall(m3ha-')and peakrainfall(maxi-
<br /> across our study region.We also include herbivores and subdivisions mum3-day rainfall total,m3ha-')at100 m resolution.Daily rainfall data
<br /> by feeding guilds that represent important indicators of resilience were generated following refs.75,76.Rainfall from each rain station
<br /> on coral reefsso,6,61.Browsers are defined as herbivores that feed on was used to derive interpolated surfaces at annual time steps using
<br /> macroalgae and associated epiphytic material,and are important for Empirical Bayesian Kriging in ArcGIS.Subwatershed catchment data'
<br /> reducing the cover of larger,more established macroalgae.Grazers were clipped to 0-10 km from the coast and used to calculate rainfall
<br /> are herbivores that feed largely on small algal turfs,helping to prevent per drainage area(Supplementary Figs.21 and 22).
<br /> their succession into larger macroalgae,and scrapers are herbivores
<br /> that closely crop the substrate and open up new space to promote the Sediment input.The Integrated Valuation of Ecosystem Services
<br /> settlement,growth and survival ofcrustose coral line algae and corals30. and Tradeoffs sediment delivery model was used to derive long-term
<br /> We followed established methods for calculating fish biomass56. annual average sediment input(kg ha-')reaching the coast77"oat100 m
<br /> The biomass of individual fishes was estimated using the allometric resolution.We then modulated the long-term annual average sedi-
<br /> length-weight conversion:W=aTLb,where parameters a and bare mentovertimeby watershed onthebasisof discharge calculated from
<br /> species-specific constants,TL is total length(cm)and Wis weight(g). peak rainfall data(Rainfall section above).Discharge bywatershed was
<br />
|