An algorithm for optically-deriving water depth from multispectral imagery in coral reef landscapes in the absence of ground-truth data

Kerr, Jeremy; Purkis, Samuel J.

doi:10.1016/j.rse.2018.03.024

Cited by 94 publications

(62 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…While correction schemes by Gordon (1997), Chavez (1988), and others effectively reduce atmospheric effects, the physics-based algorithms typically are unable to achieve depth estimates with root-mean-square errors less than ∼1-2 m in shallow waters ≤20 m (Adler-Golden et al, 2005;Lyzenga et al, 2006). Kerr and Purkis (2018) provide a bridge between the empirical and physics-based approaches, coupling forward modeling of the water column (Lee et al, 1999) with the ratio algorithm of Stumpf et al (2003) that is widely applied in empirical studies. The model of Kerr and Purkis (2018) enables the estimation of water depths up to 15 m in tropical carbonate environments without the need for in situ water depth measurements as calibration data, an important result considering that the lack of ground truth data often is the primary obstacle for estimating bathymetry in shallow-water environments.…”

Section: 1029/2018ea000539mentioning

confidence: 99%

“…Kerr and Purkis (2018) provide a bridge between the empirical and physics-based approaches, coupling forward modeling of the water column (Lee et al, 1999) with the ratio algorithm of Stumpf et al (2003) that is widely applied in empirical studies. The model of Kerr and Purkis (2018) enables the estimation of water depths up to 15 m in tropical carbonate environments without the need for in situ water depth measurements as calibration data, an important result considering that the lack of ground truth data often is the primary obstacle for estimating bathymetry in shallow-water environments. An application of the Kerr and Purkis (2018) method to RapidEye imagery from Andros Island yields a root-mean-squared error of of 1.43 m in regions shallower than 15 m (Kerr & Purkis, 2018).…”

Section: 1029/2018ea000539mentioning

confidence: 99%

“…However, our application of parsing the relationship between biosedimentary facies and water depth demands elevation precision on the order of tens of centimeters rather than meters (Maloof & Grotzinger, ). Moreover, our extensive set of ground truth data (Figure ) affords a more precise model than the robust but relatively coarse model of Kerr and Purkis ().…”

Section: Introductionmentioning

confidence: 99%

“…Our cluster‐based regression (CBR) algorithm first segments the satellite image into zones of spectral homogeneity, and then calibrates the log‐linear color‐to‐depth relationship separately for each class. The CBR algorithm applies an extension of the Stumpf et al () ratio method that uses multiple band ratios (rather than just one) to leverage more information from the satellite image (Kerr & Purkis, ). The CBR model reduces bias and error, quantifies depth uncertainty, diminishes the geographic structure of the residual, and enables more accurate estimation of water depth over a wide variety of bottom types.…”

Section: Introductionmentioning

confidence: 99%

“…All algorithms for extracting water depth from satellite imagery are based on the simple physical principle that water attenuates light. Broadly, bathymetric algorithms can be grouped into two categories: empirical methods, which use direct observations of water depth in the study area to calibrate the reflectance-to-depth relationship, and physics-based inversion algorithms that use radiative transfer models to solve for water depth without in situ calibration data (Adler-Golden et al, 2005;Albert & Gege, 2006;Brando et al, 2009;Kerr & Purkis, 2018;Lee et al, 1998Lee et al, , 1999Maritorena et al, 1994).…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A Simple Method for Extracting Water Depth From Multispectral Satellite Imagery in Regions of Variable Bottom Type

Geyman

Maloof

2019

Earth and Space Science

View full text Add to dashboard Cite

Satellite imagery offers an efficient and cost‐effective means of estimating water depth in shallow environments. However, traditional empirical algorithms for calculating water depth often are unable to account for varying bottom reflectance, and therefore yield biased estimates for certain benthic environments. We present a simple method that is grounded in the physics of radiative transfer in seawater, but made more robust through the calibration of individual color‐to‐depth relationships for separate spectral classes. Our cluster‐based regression (CBR) algorithm, applied to a portion of the Great Bahama Bank, drastically reduces the geographic structure in the residual and has a mean absolute error of 0.19 m with quantified uncertainties. Our CBR bathymetry is 3–5 times more accurate than existing models and outperforms machine learning protocols at extrapolating beyond the calibration data. Finally, we demonstrate how comparison of CBR with traditional models sensitive to bottom type reveals the characteristic length scales of biosedimentary facies belts.

show abstract

Section: 1029/2018ea000539mentioning

confidence: 99%

Section: 1029/2018ea000539mentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A Simple Method for Extracting Water Depth From Multispectral Satellite Imagery in Regions of Variable Bottom Type

Geyman

Maloof

2019

Earth and Space Science

View full text Add to dashboard Cite

show abstract

Mapping Apennines river paths along different hydrological conditions from satellite images: A description of the method and potential applications

Giacomelli

Chelli

Bresciani

et al. 2023

River Research & Apps

View full text Add to dashboard Cite

Climate change is producing large impacts on rivers, amplifying hydrological extremes. Prolonged drought periods result in dramatic stress for river biota and associated processes due to low discharge, reducing the interactions between rivers and their lateral environments or leading to hydrological intermittency. New quantitative methods are needed, to correlate discharge with the available riverine habitats. In this work we have mapped the wet surface and paths of two stretches of the Taro and Trebbia Apennine rivers, analyzing satellite images from periods with contrasting discharge. The considered stretches are critical due to different human pressures (large water withdrawals for agriculture and industrial use) and are particularly vulnerable to further, climate‐driven discharge reductions. The produced images offer multiple possibilities to extract qualitative and quantitative information at the whole stretch scales, including habitat reduction along with decreasing discharge, threshold discharge limiting lateral interactions, or the evaluation of longitudinal river continuity. We discuss the limitation and the potentialities of the method and the maps produced in terms of possible application in the field of river geomorphology, ecology, the definition of ecological river flow, risk assessment, and river management.

show abstract

High-resolution habitat and bathymetry maps for 65,000 sq. km of Earth’s remotest coral reefs

Purkis

Gleason

Purkis³

et al. 2019

Coral Reefs

Self Cite

View full text Add to dashboard Cite

With compelling evidence that half the world's coral reefs have been lost over the last four decades, there is urgent motivation to understand where reefs are located and their health. Without such basic baseline information, it is challenging to mount a response to the reef crisis on the global scale at which it is occurring. To combat this lack of baseline data, the Khaled bin Sultan Living Oceans Foundation embarked on a 10-yr survey of a broad selection of Earth's remotest reef sites-the Global Reef Expedition. This paper focuses on one output of this expedition, which is meter-resolution seafloor habitat and bathymetry maps developed from DigitalGlobe satellite imagery and calibrated by field observations. Distributed on an equatorial transect across 11 countries, these maps cover 65,000 sq. km of shallow-water reef-dominated habitat. The study represents an order of magnitude greater area than has been mapped previously at high resolution. We present a workflow demonstrating that DigitalGlobe imagery can be processed to useful products for reef conservation at regional to global scale. We further emphasize that the performance of our mapping workflow does not deteriorate with increasing size of the site mapped. Whereas our workflow can produce regional-scale benthic habitat maps for the morphologically diverse reefs of the Pacific and Indian oceans, as well as the more depauperate reefs of the Atlantic, accuracies are substantially higher for the former than the latter. It is our hope that the map products delivered to the community by the Living Oceans Foundation will be utilized for conservation and act to catalyze new initiatives to chart the status of coral reefs globally.

show abstract

An algorithm for optically-deriving water depth from multispectral imagery in coral reef landscapes in the absence of ground-truth data

Cited by 94 publications

References 43 publications

A Simple Method for Extracting Water Depth From Multispectral Satellite Imagery in Regions of Variable Bottom Type

A Simple Method for Extracting Water Depth From Multispectral Satellite Imagery in Regions of Variable Bottom Type

Mapping Apennines river paths along different hydrological conditions from satellite images: A description of the method and potential applications

High-resolution habitat and bathymetry maps for 65,000 sq. km of Earth’s remotest coral reefs

Contact Info

Product

Resources

About