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

Geyman, Emily C.; Maloof, Adam C.

doi:10.1029/2018ea000539

Cited by 82 publications

(60 citation statements)

References 60 publications

(101 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…To calibrate the conversion from satellite image color to water depth (see section ), we use

\sim

50,000 depth measurements collected with a BioSonics MX Aquatic Habitat Echosounder (single frequency, 204.8 kHz) (Geyman & Maloof, ).…”

Section: Methodsmentioning

confidence: 99%

“…(A) depicts the workflow if the user has local bathymetric data. (1A) Use a bathymetric model (e.g., Geyman & Maloof, ), calibrated by local water depth measurements, to compute an instantaneous water depth map from each satellite image. (2A) Use each satellite image's time stamp to compute the hours from high tide at the reference tide station.…”

Section: Methodsmentioning

confidence: 99%

“…We use a simple modification of the Lyzenga () linear algorithm in which we segment the satellite image into zones of spectral homogeneity and calibrate different color‐to‐depth relationships for each zone before applying a fuzzy

k

‐nearest neighbor fit to solve for the final bathymetry. This cluster‐based regression (CBR) method of Geyman and Maloof () reduces error by

\geq

40% relative to previous bathymetric estimates for the region in Figure (Geyman & Maloof, ), since it better accounts for variability in the spectral properties of the sea floor. When applied to the 56 Landsat images in our compilation, the CBR method produces depth estimates with mean absolute errors (MAEs) of 0.14–0.26 m and

R^{2}

values of 0.70–0.88 (Figure ; Table S1).…”

Section: Methodsmentioning

confidence: 99%

“…Example model results for the cluster‐based regression (CBR) bathymetry algorithm (Geyman & Maloof, ) applied to a Landsat 8 image acquired on 20 March 2018. The “true depth” (

x

axis) represents tide‐corrected echosounder depth measurements acquired within

\leq

5 km of a tide gauge (Figure ), and the “model depth” (

y

axis) represents the water depth calculated from the closest pixel in the Landsat image (30‐m resolution).…”

Section: Methodsmentioning

confidence: 99%

“…“TOTO” stands for “Tongue of the Ocean,” the

\sim

1‐ to 2‐km‐deep trough east of Andros Island. The gray lines represent the positions of

\sim

300,000 echosounder depth measurements (Geyman & Maloof, ). Only the

\sim

50,000 depth measurements acquired within

\leq

5 km of a tide gauge, where our understanding of the tides is well constrained, are used as calibration data.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Deriving Tidal Structure From Satellite Image Time Series

Geyman

Maloof

2020

Earth and Space Science

Self Cite

View full text Add to dashboard Cite

In shallow coastal regions, tides often control the water flux, which in turn directs sediment transport, nutrient delivery, and geochemical gradients. However, tides in shallow areas are spatially heterogeneous, making it challenging to constrain the geographic structure of tidal phase and amplitude without extensive networks of tide gauges. We present a simple remote sensing method for deriving tidal structure from satellite time series. Our method is based on two observations: (1) Tidally driven variations in water depth can be detected as changes in pixel intensity in optical satellite imagery, and (2) repeating passes by an orbiting satellite capture a region at different phases of the tidal cycle. By stacking multiple satellite acquisitions of a shallow bank, we can compute the relative tidal phase and amplitude for each pixel location, thereby resolving a detailed map of tidal propagation and attenuation. While our method requires a set of local water‐depth measurements to calibrate the color‐to‐depth relationship and compute tidal amplitude (in meters), our method can produce spatial estimates of tidal phase and relative amplitude without any site‐specific calibration data. As an illustration of the method, we use Landsat imagery to derive the spatial structure of tides on the Great Bahama Bank, estimating tidal phase and amplitude with mean absolute errors of 15 min and 0.15 m, respectively.

show abstract

“…To calibrate the conversion from satellite image color to water depth (see section ), we use

\sim

50,000 depth measurements collected with a BioSonics MX Aquatic Habitat Echosounder (single frequency, 204.8 kHz) (Geyman & Maloof, ).…”

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

k

‐nearest neighbor fit to solve for the final bathymetry. This cluster‐based regression (CBR) method of Geyman and Maloof () reduces error by

\geq

R^{2}

values of 0.70–0.88 (Figure ; Table S1).…”

Section: Methodsmentioning

confidence: 99%

“…Example model results for the cluster‐based regression (CBR) bathymetry algorithm (Geyman & Maloof, ) applied to a Landsat 8 image acquired on 20 March 2018. The “true depth” (

x

axis) represents tide‐corrected echosounder depth measurements acquired within

\leq

5 km of a tide gauge (Figure ), and the “model depth” (

y

axis) represents the water depth calculated from the closest pixel in the Landsat image (30‐m resolution).…”

Section: Methodsmentioning

confidence: 99%

“…“TOTO” stands for “Tongue of the Ocean,” the

\sim

1‐ to 2‐km‐deep trough east of Andros Island. The gray lines represent the positions of

\sim

300,000 echosounder depth measurements (Geyman & Maloof, ). Only the

\sim

50,000 depth measurements acquired within

\leq

5 km of a tide gauge, where our understanding of the tides is well constrained, are used as calibration data.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Deriving Tidal Structure From Satellite Image Time Series

Geyman

Maloof

2020

Earth and Space Science

Self Cite

View full text Add to dashboard Cite

show abstract

The distribution of depth, volume, and basin shape for lakes in the conterminous United States

Ganz,

Glines,

Rose

2023

Limnology & Oceanography

View full text Add to dashboard Cite

Depth regulates many attributes of aquatic ecosystems, but relatively few lakes are measured, and existing datasets are biased toward large lakes. To address this, we used a large dataset of maximum (Zmax; n = 16,831) and mean (Zmean; n = 5,881) depth observations to create new depth models, focusing on lakes < 1,000 ha. We then used the models to characterize patterns in lake basin shape and volume. We included terrain metrics, water temperature and reflectance, polygon attributes, and other predictors in a random forest model. Our final models generally outperformed existing models (Zmax ; root mean square error [RMSE] = 8.0 m and Zmean ; RMSE = 3.0 m). Our models show that lake depth followed a Pareto distribution, with 2.8 orders of magnitude fewer lakes for an order of magnitude increase in depth. In addition, despite orders of magnitude variation in surface area, most size classes had a modal maximum depth of ~ 5 m. Concave (bowl‐shaped) lake basins represented 79% of all lakes, but lakes were more convex (funnel‐shaped) as surface area increased. Across the conterminous United States, 9.8% of all lake water was within the top meter of the water column, and 48% in the top 10 m. Excluding the Laurentian Great Lakes, we estimate the total volume in the conterminous United States is 1,057–1,294 km3, depending on whether Zmax or Zmean was modeled. Lake volume also exhibited substantial geographic variation, with high volumes in the upper Midwest, Northeast, and Florida and low volumes in the southwestern United States.

show abstract