Assessment of Sentinel-2 and Landsat-8 OLI for Small-Scale Inland Water Quality Modeling and Monitoring Based on Handheld Hyperspectral Ground Truthing

Abstract

This study investigates the best available methods for remote monitoring inland small-scale waterbodies, using remote sensing data from both Landsat-8 and Sentinel-2 satellites, utilizing a handheld hyperspectral device for ground truthing. Monitoring was conducted to evaluate water quality indicators: chlorophyll-a (Chl-a), colored dissolved organic matter (CDOM), and turbidity. Ground truthing was performed to select the most suitable atmospheric correction technique (ACT). Several ACT have been tested: dark spectrum fitting (DSF), dark object subtraction (DOS), atmospheric and topographic correction (ATCOR), and exponential extrapolation (EXP). Classical sampling was conducted first; then, the resulting concentrations were compared to those obtained using remote sensing analysis by the above-mentioned ACT. This research revealed that DOS and DSF achieved the best performance (an advantage ranging between 29% and 47%). Further, we demonstrated the appropriateness of the use of Sentinel-2 red and vegetation red edge reciprocal bands <a:math xmlns:a="http://www.w3.org/1998/Math/MathML" id="M1"> <a:mfenced open="(" close=")"> <a:mrow> <a:mn>1</a:mn> <a:mo>/</a:mo> <a:mfenced open="(" close=")"> <a:mrow> <a:mtext>B</a:mtext> <a:mn>4</a:mn> <a:mo>×</a:mo> <a:mtext>B</a:mtext> <a:mn>6</a:mn> </a:mrow> </a:mfenced> </a:mrow> </a:mfenced> </a:math> for estimating Chl-a ( <g:math xmlns:g="http://www.w3.org/1998/Math/MathML" id="M2"> <g:msup> <g:mrow> <g:mi>R</g:mi> </g:mrow> <g:mrow> <g:mn>2</g:mn> </g:mrow> </g:msup> <g:mo>=</g:mo> <g:mn>0.82</g:mn> </g:math> , <i:math xmlns:i="http://www.w3.org/1998/Math/MathML" id="M3"> <i:mtext>RMSE</i:mtext> <i:mo>=</i:mo> <i:mn>14.52</i:mn> <i:mtext> </i:mtext> <i:mtext>mg</i:mtext> <i:mo>/</i:mo> <i:msup> <i:mrow> <i:mtext>m</i:mtext> </i:mrow> <i:mrow> <i:mn>3</i:mn> </i:mrow> </i:msup> </i:math> ). As for Landsat-8, red to near-infrared ratio ( <k:math xmlns:k="http://www.w3.org/1998/Math/MathML" id="M4"> <k:mtext>B</k:mtext> <k:mn>4</k:mn> <k:mo>/</k:mo> <k:mtext>B</k:mtext> <k:mn>5</k:mn> </k:math> ) produced the best performing model ( <m:math xmlns:m="http://www.w3.org/1998/Math/MathML" id="M5"> <m:msup> <m:mrow> <m:mi>R</m:mi> </m:mrow> <m:mrow> <m:mn>2</m:mn> </m:mrow> </m:msup> <m:mo>=</m:mo> <m:mn>0.71</m:mn> </m:math> , <o:math xmlns:o="http://www.w3.org/1998/Math/MathML" id="M6"> <o:mtext>RMSE</o:mtext> <o:mo>=</o:mo> <o:mn>39.88</o:mn> <o:mtext> </o:mtext> <o:mtext>mg</o:mtext> <o:mo>/</o:mo> <o:msup> <o:mrow> <o:mtext>m</o:mtext> </o:mrow> <o:mrow> <o:mn>3</o:mn> </o:mrow> </o:msup> </o:math> ), but it did not perform as well as Sentinel-2. Regarding turbidity, the best model ( <q:math xmlns:q="http://www.w3.org/1998/Math/MathML" id="M7"> <q:msup> <q:mrow> <q:mi>R</q:mi> </q:mrow> <q:mrow> <q:mn>2</q:mn> </q:mrow> </q:msup> <q:mo>=</q:mo> <q:mn>0.85</q:mn> </q:math> , <s:math xmlns:s="http://www.w3.org/1998/Math/MathML" id="M8"> <s:mtext>RMSE</s:mtext> <s:mo>=</s:mo> <s:mn>0.87</s:mn> </s:math> NTU) obtained by Sentinel-2 utilized a single band (B4), while the best model (with <u:math xmlns:u="http://www.w3.org/1998/Math/MathML" id="M9"> <u:msup> <u:mrow> <u:mi>R</u:mi> </u:mrow> <u:mrow> <u:mn>2</u:mn> </u:mrow> </u:msup> <u:mo>=</u:mo> <u:mn>0.64</u:mn> </u:math> , <w:math xmlns:w="http://www.w3.org/1998/Math/MathML" id="M10"> <w:mtext>RMSE</w:mtext> <w:mo>=</w:mo> <w:mn>0.90</w:mn> </w:math> NTU) using Landsat-8 was performed by applying two bands ( <y:math xmlns:y="http://www.w3.org/1998/Math/MathML" id="M11"> <y:mtext>B</y:mtext> <y:mn>1</y:mn> <y:mo>/</y:mo> <y:mtext>B</y:mtext> <y:mn>3</y:mn> </y:math> ). Mapping the water quality parameters using the best performance biooptical model showed the significant effect of the adjacent land on the boundary pixels compared to pixels of deeper water.

Topics & Concepts