Consistency of Landsat-8/-9 reflectance products for aquatic science and applications
Sakib Kabir, SSAI / NASA GSFC, sakib.kabir@ssaihq.com (Presenter)
Nima Pahlevan, SSAI / NASA GSFC, nima.pahlevan@nasa.gov
Ryan O'Shea, SSAI / NASA GSFC, ryan.e.o'shea@nasa.gov
Brian B. Barnes, University of South Florida, bbarnes4@usf.edu
The Landsat-9 (L9) Operational Land Imager 2 (OLI2) is the continuity of the Landsat program and largely identical to Landsat-8 (L8) Operational Land Imager (OLI). Even though the Landsat sensors are designed for terrestrial science and applications, high-quality OLI imagery have been used extensively to study and monitor aquatic ecosystems (e.g., rivers, lakes, reservoirs, coastal estuaries, and shelves). OLI’s success in monitoring aquatic ecosystems stems from its enhanced radiometric quality. OLI2 observations are anticipated to provide equivalent or better performance than OLI in every domain and complement existing missions for their synergistic use. To exploit the OLI2 observations, the radiometric accuracy/precision of OLI2 should be evaluated and quantified over different water bodies, such that its performance is well characterized and quantified for relevant studies. During the commissioning phase, L9 was nominally operated under L8 on the 11th to 17th of November 2021, allowing for near-simultaneous observations. This study leverages those near-coincident OLI-OLI2 observations (~ 2-minute time difference) and provides relative consistency evaluation of the standard United States Geological Survey (USGS) top-of-atmosphere (TOA) reflectance (ρ_(t )) and atmospherically corrected reflectance (aquatic reflectance; ρ_w^AR) products over various types of water bodies (e.g., coastal, inland, and open ocean). Overall, OLI/OLI2 ρ_(t ) were found within 0.3% for the visible and infrared bands, except green band showing ~1.3% difference. The median ρ_w^AR differences were ~0.7 – 4% for the visible bands. During the underfly maneuver, OLI2 and OLI scene pairs overlapped from a few percent to 100%, providing a unique dataset to analyze the sun-sensor geometry correction performance of atmospheric correction (AC) processors while generating ρ_(w ). Hence, we compared three widely used AC methods, namely SeaDAS, ACOLITE, and POLYMER, for their rigor in correcting geometric effects. Overall, SeaDAS-derived ρ_(w )were found most consistent. Additionally, the signal-to-noise ratio (SNR) of OLI2 and OLI were intercompared over coastal and inland water for the visible and infrared bands. OLI2 SNRs were found 7 – 20% higher than OLI for the visible bands, likely due to its 14-bit quantization rate.
Poster: Poster_Kabir_3-47_229_35.pdf
Associated Project(s):
Poster Location ID: 3-47
Presentation Type: Poster
Session: Poster Session 3
Session Date: Thu (May 11) 3:00-5:00 PM
CCE Program: OBB