Validation, trend analysis & bias reduction through satellite fusion of the cloud-top height record from terra MODIS & MISR

Mitra, Arka

Permalink

https://hdl.handle.net/2142/120492 Copy

Description

Title

Validation, trend analysis & bias reduction through satellite fusion of the cloud-top height record from terra MODIS & MISR

Author(s)

Mitra, Arka

Issue Date

2023-02-22

Director of Research (if dissertation) or Advisor (if thesis)

Di Girolamo, Larry

Doctoral Committee Chair(s)

Di Girolamo, Larry

Committee Member(s)

• Nesbitt, Stephen
• Sriver, Ryan
• Proistosescu, Cristian

Department of Study

Atmospheric Sciences

Discipline

Atmospheric Sciences

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Cloud-Top Height
• NASA Terra
• Satellite Cloud Record
• Climate Trends
• Satellite Fusion

Abstract

Modeled clouds contribute the largest uncertainty in modeled responses to climate change. While significant uncertainties exist in predictions of the complicated pathways in which cloud amounts (CA), cloud-top heights (CTH) and cloud opacity respond to warmer temperatures, there is robust agreement between qualitative aspects of cloud and climate responses in models, such as rising high-topped clouds and expanding subtropical highs. Since the response in global CTH is the most robust of such responses, accurate estimation of long-term variability in the vertical distribution of CTH on a global scale is important for reducing climate modeling uncertainty and understanding long-term variability in cloud properties. Such analyses require accurate and precise climate records of global cloud property retrievals from satellite sensors that maintain long-term orbital stability. While recent space-based active sensor records have improved our understanding of vertical cloud distributions, satellite passive sensors are better suited for detecting trends due to their longer records and broad swaths yielding greater global coverage. Our longest records of global CTH from passive sensors in a stable orbit come from the 22-year (2000-2021) record of the Multiangle Imaging Spectroradiometer (MISR) and the Moderate Resolution Imaging Spectroradiometer (MODIS), on board the Terra satellite. The stable orbit ensured that Terra MODIS and MISR climate records were not aliased by diurnal and other spurious modes of variability that could otherwise obfuscate the estimation of long-term climate trends. Thus, to ensure the robustness of the Terra record for current and future studies on the Earth’s climate, we quantified the MISR and MODIS CTH error characteristics vis-à-vis cloud geometrical and optical properties, through the first direct, semi-global comparison against a space-based lidar (ISS-CATS). We establish that MISR and MODIS records of CTH are highly accurate and precise for climate analyses, with their mean errors over single-layered, unbroken, optically thick clouds at all atmospheric levels being −280 ± 370 m and −540 ± 690 m, respectively. For the first time, we were able to close the MISR CTH error budget by decomposing the net MISR CTH bias and random errors into constituent errors, such as altitude-dependent MISR wind-retrieval errors and the very first estimate of a stereo-opacity bias (i.e., the retrieval of stereo heights within a depth of the cloud when the extinction near cloud-top is low). On the other hand, MODIS CTH errors are largely driven by CTH bias errors for geometrically thick cirrus and random errors due to forward modeling uncertainty. Furthermore, we established that scenes with 2-layered cloud systems with thin cirrus over thicker low clouds lead to significant disagreements between the two sensors. MISR detects the lower cloud in 2-layered systems if cirrus optical depth <∼0.3, but MISR low-cloud CTH errors are unaltered by the presence of thin cirrus. Meanwhile, for such 2-layered systems, MODIS underestimates top-layer CTH by greater than 1 km for cirrus optical depths < 0.8, thus producing more midlevel clouds than MISR. With the uncertainties in MISR and MODIS CTH records better quantified, we studied long-term variability (over the Terra climate record) in MISR and Terra-MODIS CA as function of cloud altitude and cloud opacity. MISR and MODIS high cloud amounts significantly increased between 2000-2021 by ~0.08-0.2% about the climatological tropopause, whereas their total cloud fractions significantly decreased by ~0.6-2.0% over the subtropical oceans, leading to significant reduction in shortwave albedo. These trends demonstrate for the first time that the predicted rise of high-topped clouds is statistically significant over all latitudes between 60ºN-60ºS from both MISR and MODIS, while lending further evidence for subtropical expansion during the Terra record. In spite of employing different retrieval techniques, there is broad agreement between the large-scale spatio-temporal variation of height-resolved and total CA from MODIS and MISR over regions dominated by single-layered clouds. However, the inability of passive sensors to separate CA, CTH and opacity changes for multi-layered cloud systems likely leads to opposite trends in vertically integrated high and low cloud amounts and total cloud fractions over tropical oceans. Since both the validation against ISS-CATS lidar and trend analysis of the Terra record demonstrate that applying single cloud-layer assumptions to multi-layered cloud systems lead to large disagreements between MISR and MODIS pixel-level CTH retrievals and long-term trends, we conceptualized and implemented a MISR+MODIS fusion algorithm to improve the accuracy of Terra high-cloud CTH and emissivities in 2-layered scenes. This is relevant because multi-layered clouds constitute ~30% of global cloud cover (out of which ~80% constitute cirrus over lower-level water or mixed-phase clouds). Through comparison against collocated ISS-CATS data, we show that our new algorithm improves the accuracies in high cloud CTH and emissivity by ~75% as compared to standard MODIS products. This leads to significant improvements to our estimates of modeled longwave radiative effect of clouds in multi-layered scenes (between 5-45 W m-2, depending on the macrophysical and optical properties of the cloud layers). Owing to this large radiative impact, we strongly suggest that future work scale up the pixel-level MISR+MODIS fusion algorithm over the entire MISR swath (i.e., over the domain where MODIS and MISR make concurrent observations) to derive the first distributions of 2-layered cloud properties from the morning orbit of Terra. Such a climate record will be crucial in determining better estimates of long-term variability in vertically resolved cloud properties, which can lead to better observational constraints on modeled cloud feedback uncertainties. These advances made here may also inform and improve future satellite missions seeking to observe cloud properties from space.

Graduation Semester

2023-05

Type of Resource

Thesis

Handle URL

https://hdl.handle.net/2142/120492

Copyright and License Information

Chapter 2 is a paper published in the Journal of Geophysical Research – Atmospheres (Mitra et al., 2021) with dissertation format. This paper is available with open access at: https://doi.org/10.1029/2020JD034281. This paper is distributed under the Creative Commons Attribution 4.0 License (https://creativecommons.org/licenses/by/4.0/) by the Wiley Online Library (https://onlinelibrary.wiley.com). Under this license, the document can be shared and adapted provided that appropriate credit has been given to the authors, the license has been linked, and changes (if any) have been indicated. For more information on this journal’s policy regarding license and copyright please read: https://authorservices.wiley.com/author-resources/Journal-Authors/licensing/index.html. Chapter 4 is a paper currently in review in the Journal of Geophysical Research – Atmospheres with dissertation format. Appendix A is from the supplement of this manuscript. The current version of the submitted manuscript along with its supplement (Mitra et al., 2022) is available to read in pre-print format with open access at: https://doi.org/10.1002/essoar.10512795.1. This preprint is distributed under the Attribution-NoDerivs 2.0 Generic (CC BY-ND 2.0) License (https://creativecommons.org/licenses/by-nd/2.0/) by the ESS Open Archive (https://essopenarchive.org/). Under this license, the document can be shared and adapted provided that appropriate credit has been given to the authors, the license has been linked, and changes (if any) have been indicated. For more information on this journal’s policy regarding license and copyright please read: https://essopenarchive.org/users/3/articles/586003-faqs.

Owning Collections