Content uploaded by Thomas Plocoste
Author content
All content in this area was uploaded by Thomas Plocoste on Oct 22, 2024
Content may be subject to copyright.
7th Indian International Conference on Air Quality Management (IICAQM-2022)
IIT Madras, 29 November - 1 December 2022
MULTISCALE ANALYSIS OF PM2.5-PM10 RELATIONSHIPS IN THE
CARIBBEAN AREA
T. Plocoste1, 2, S. Adarsh3*
1Department of Research in Geoscience, KaruSphère SASU, Abymes 97139, Guadeloupe
(F.W.I.), France
2Univ Antilles, LaRGE Laboratoire de Recherche en Géosciences et Energies (EA 4539), F-
97100 Pointe-à-Pitre, France
3*TKM College of Engineering Kollam, Kerala, India
Tel.: +91 8078976388, E-mail: adarsh_lce@yahoo.co.in
ABSTRACT: This paper presents Hilbert Huang Transform (HHT) based approaches of Time
Dependent Intrinsic Correlation and Time Dependent Intrinsic Cross Correlation approaches for
investigating the relationships between PM2.5 and PM10. The study uses datasets of Puerto-Rico
between 2007 and 2010 period for demonstrating the relationships between two prime particulate
matter concentration datasets of air pollution in multiple time scales. The method first decomposes
both the series using Improved CEEMDAN to get the periodic scales. The Time Dependent Intrinsic
Correlation (TDIC) method was invoked between the resulting modes with similar periodic scales.
The TDIC method showed that for the correlation pattern is strongest and stable at seasonal and annual
cycles. Subsequently the lagged correlations at each time scales were analysed using TDI cross
correlation method. In the case of high frequency IMFs of PM10 less than seasonal scale, the value of
the IMF at a generic time scale is dependent on multiple antecedent values of PM2.5. However
seasonal scale onwards the correlation pattern of PM2.5-PM10 relationship is stable and IMFs of
PM10 at these scales can be modelled by lag1 IMF of PM2.5.
Keywords: Multiscale, Correlation, PM10, PM2.5, Caribbean area