1,2,1- 1Georg-August-Universität Göttingen, Faculty of Geosciences and Geography, Department of Physical Geography, Germany
- 2Georg-August-Universität Göttingen, Faculty of Geosciences and Geography, Department of Cartography, GIS and Remote sensing, Germany
MODIS land surface temperature (LST) data is widely used in climate studies, including assessing urban heat intensity. However, its quality varies across space and time, especially in complex and rapidly urbanizing landscapes. Yet, seasonal variation in data quality is rarely reported for annual data, partly due to the tediousness in decoding and interpreting data quality.
To simplify reporting, we propose two quality indices for MODIS daily LST data: (1) Usefulness index (UI) quantifies pixel level data reliability in two dimensions, i.e., spatial (sUI), as proportion of reliable pixels across an area of interest for a given day, and temporal (tUI), as frequency of reliable pixels for a given location over time (e.g., monthly); (2) Balance index (BI) measures variation in tUI for a given period (e.g. annually). Both indices range from 0 to 100%, with higher UI indicating greater reliability and lower BI reflecting uniform data representation. Reliable pixels are identified using the QC band of MODIS, where 8-bit quality information is simplified to a single measure of pixel reliability.
We tested the indices in Kumasi, Ghana (tropical) and Shanghai, China (subtropical), using MODIS Aqua and Terra data from 2000 to 2022 across four daily observational times. For Kumasi, mean sUI was 32.9%, tUI was 20.2% (~6 reliable days per month), and BI was 76.2%, indicating low spatial and temporal reliability, dominated by four dry-season months: November to February. In Shanghai, mean sUI was 39.3%, tUI was 25.7%, and BI was 46.8%, suggesting slightly higher reliability and better monthly representation, with June as the least reliable month.
Our findings emphasize the need for rigorous quality reporting of annual MODIS-LST data in tropical regions with pronounced seasonal variability, such as Kumasi. We advocate routine use of UI and BI to improve reliability and comparability of aggregated LST data in climate studies.
How to cite: Li, L., Sauer, D., Putzenlechner, B., and Asabere, S. B.: Usefulness and balance indices; novel metrics for assessing MODIS–land surface temperature quality in annual aggregation , 12th International Conference on Urban Climate, Rotterdam, The Netherlands, 7–11 Jul 2025, ICUC12-222, https://doi.org/10.5194/icuc12-222, 2025.
Supporters & sponsors
