Work

Remote Sensing of Lake Water Clarity: Performance and Transferability of Both Historical Algorithms and Machine Learning

Year: 2021

Type: article

Abstract: There has been little rigorous investigation of the transferability of existing empirical water clarity models developed at one location or time to other lakes and dates of imagery with differing conditions. Machine learning methods have not been widely adopted for analysis of lake optical properties such as water clarity, despite their successful use in many other applications of environmental remote sensing. This study compares model performance for a random forest (RF) machine learning algorithm and a simple 4-band linear model with 13 previously published empirical non-machine learning algorithms. We use Landsat surface reflectance product data aligned with spatially and temporally co-located in situ Secchi depth observations from northeastern USA lakes over a 34-year period in this analysis. To evaluate the transferability of models across space and time, we compare model fit using the complete dataset (all images and samples) to a single-date approach, in which separate models are developed for each date of Landsat imagery with more than 75 field samples. On average, the single-date models for all algorithms had lower mean absolute errors (MAE) and root mean squared errors (RMSE) than the models fit to the complete dataset. The RF model had the highest pseudo-R2 for the single-date approach as well as the complete dataset, suggesting that an RF approach outperforms traditional linear regression-based algorithms when modeling lake water clarity using satellite imagery.

Source: Remote Sensing

DOI: 10.3390/rs13081434

Authors: Hannah J. Rubin, David A. Lutz, Bethel G. Steele, Kathryn L. Cottingham, Kathleen C. Weathers, Mark J. Ducey, Michael Palace, Kenneth M. Johnson, Jonathan W. Chipman

Institutions: Dartmouth College, University of Tennessee at Knoxville, Cary Institute of Ecosystem Studies, University of New Hampshire

Language: English

Topic: Aquatic Ecosystems and Phytoplankton Dynamics

Subfield: Environmental Chemistry

Field: Environmental Science

Domain: Physical Sciences

SDG: Clean water and sanitation

Open Access status: gold

Estimated APC paid: $2,592

Funders: National Aeronautics and Space Administration, Dartmouth College

Award: HUMAN SOCIETIES RELY ON LAKES AND RESERVOIRS FOR MANY ESSENTIAL ECOSYSTEM SERVICES INCLUDING THE SUPPLY OF DRINKING WATER OPPORTUNITIES FOR RECREATION AESTHETIC BENEFITS AND HABITAT FOR BIODIVERSITY. YET HUMANS ALSO IMPACT THE PROVISIONING OF THESE SERVICES BY ALTERING THE LANDSCAPE WITHIN LAKE CATCHMENTS. SUCH CHANGES FOR INSTANCE THE DEVELOPMENT OF BUILT STRUCTURES AND ROADS TYPICALLY RESULT IN THE LOSS OF NATURAL FOREST AND VEGETATIVE COVER AN INCREASE IN IMPERVIOUS SURFACES AND HABITAT FRAGMENTATION ALL OF WHICH INCREASE NUTRIENT LOADING TO LAKES AND DECREASE WATER QUALITY THROUGH THE PROMOTION OF EXCESSIVE PHYTOPLANKTON GROWTH. CYANOBACTERIAL BLOOMS ONE OF THE MAJOR CONSEQUENCES OF SUCH EXCESSIVE GROWTH CAN SIGNIFICANTLY IMPEDE THE DELIVERY OF ENVIRONMENTAL BENEFITS BECAUSE THEY DEGRADE WATER QUALITY CLARITY AND AESTHETICS STIFLE LAKE USE AND DECREASE NEARBY PROPERTY VALUES. TO DATE THE INTERACTING EFFECTS OF HUMAN AND ENVIRONMENTAL PRESSURES ON LAKE WATER QUALITY ARE NOT HOLISTICALLY UNDERSTOOD. THIS INHIBITS THE ABILITY OF STAKEHOLDERS TO MANAGE LAKES FOR RESILIENCE A CONCERN GIVEN THAT LAKES ARE FACING GROWING PRESSURE FROM BOTH LAND USE AND CLIMATE CHANGE. TO ADDRESS THIS ISSUE THE OVERARCHING GOAL OF THIS PROJECT IS TO TRANSFORM THE STATE OF KNOWLEDGE OF HOW DEMOGRAPHIC AND LAND COVER CHANGES WITHIN LAKE CATCHMENTS INTERACT WITH REGIONAL CLIMATE CHANGE AND THEREBY IMPACT LAKE WATER QUALITY PARTICULARLY WATER CLARITY AND THE INCIDENCE OF CYANOBACTERIAL BLOOMS. WE WILL FOCUS ON LAKES WITHIN PRIMARILY FORESTED WATERSHEDS IN THE NORTHEASTERN UNITED STATES PARTICULARLY THOSE THAT ARE SUBJECT TO ONGOING DEVELOPMENT AND MAJOR SHIFTS IN CLIMATE. OUR SPECIFIC OBJECTIVES ARE TO: 1.EVALUATE THE INTERACTIVE EFFECTS OF HUMAN POPULATION GROWTH ALTERED LAND USE AND LAND COVER CLIMATE VARIABILITY AND LAKE-SPECIFIC CHARACTERISTICS ON LAKE WATER QUALITY FROM 1986-2016 THEREBY DEVELOPING PREDICTIVE MODELS OF HOW LAKE WATER QUALITY MAY CHANGE IN THE FUTURE. 2.DEVELOP MUCH-NEEDED TOOLS AND MODELS FOR REAL-TIME ASSESSMENT OF CYANOBACTERIAL BLOOMS IN SMALL-TO-MODERATE-SIZED LAKES USING NEW PLATFORMS AND TECHNOLOGIES INCLUDING SUB-ORBITAL UNMANNED AERIAL SYSTEM-BASED HYPERSPECTRAL SENSORS AND CROWD-SOURCED CITIZENSCIENCE DATA COLLECTED USING SMARTPHONE APPLICATIONS. ALTOGETHER THE PROPOSED RESEARCH WILL SIGNIFICANTLY ADVANCE OUR ABILITY TO FORECAST LAKE CONDITIONS ON BOTH AN INDIVIDUAL BASIS AND ACROSS LANDSCAPES. ALTHOUGH FEW STUDIES HAVE EVALUATED HUMAN DEMOGRAPHY AS A DRIVER OF CATCHMENT CHANGE WE WILL INTEGRATE THIS IMPORTANT DISCIPLINE TO BROADEN OUR UNDERSTANDING OF LAKE ECOSYSTEMS IN AN INTERDISCIPLINARY FASHION. FURTHERMORE WE WILL USE SATELLITE AND SUB-ORBITAL REMOTELY SENSED DATA AT A VARIETY OF SPECTRAL AND SPATIAL RESOLUTIONS TO QUANTIFY CHANGES IN LAND COVER LAKE TEMPERATURE CLARITY AND ALGAL INDICATORS OVER TIME FOR CATCHMENTS AND WATER BODIES AND THEREBY WILL CONTRIBUTE TO NASA S GOALS OF COMBINING REMOTELY SENSED AND IN-LAKE MEASUREMENTS WITH OBSERVATIONAL MODELING TO PROVIDE GUIDANCE TO NATURAL RESOURCE MANAGERS. OUR PROPOSED PROJECT WILL THUS PROVIDE CRUCIAL GUIDANCE FOR THE DEVELOPMENT AND ADVANCEMENT OF CURRENT AND FUTURE NASA SENSORS PARTICULARLY IN THE HYPERSPECTRAL DOMAIN FOR EFFECTIVE MONITORING AND MANAGEMENT OF THE PLANET S FRESHWATER RESOURCES.

OpenAlex ID: W3154489171

FWCI: 1.981

Cites: 110

Cited by: 41

Related to: 10