This study investigates the relationship between particulate matter that have an aerodynamic diameter less than 10 diameter ($PM10$) and air temperature (T) in the Caribbean basin. To conduct this analysis, eleven years of daily time series recorded in Guadeloupe archipelago were used. Firstly, the stationarity threshold of both time series is quantified using the infinite moving average. Then, statistical analyses were performed using the conditional mean and the joint probability density function method. Both methods used showed a dependence between $PM10$ and T particularly during the high dust season, i.e. from May to September. From a dynamical point of view, the coherence function highlighted a global correlation information between $PM10$ and T in the frequency domain. To capture possible localized correlation information, both time series are studied in a multiscale way using a new cross-correlation technique classically based on Empirical Mode Decomposition (EMD) termed the Time-Dependent Intrinsic Correlation (TDIC) analysis. After computing EMD and Ensemble Empirical Mode Decomposition (EEMD) methods, a mode-mixing problem is detected in EMD frame. Consequently, for the first time in air pollution field, a combination of EEMD and TDIC is performed. The achieved results show that the strength and nature of association between $PM10$ and T vary with time scales and time spells. We found that $PM10$ from African and marine aerosols are strongly correlated to T respectively for large and small periodicities. One also observed that extreme weather events such as heavy rainfall, hurricanes or drought induce a strong anti-correlation.

这项研究调查了空气动力学直径小于10直径的颗粒物之间的关系（$P\mathrm{中号}10$）和加勒比海盆地的气温（T）。为了进行此分析，使用了瓜德罗普群岛记录的11年每日时间序列。首先，使用无限移动平均值对两个时间序列的平稳性阈值进行量化。然后，使用条件均值和联合概率密度函数方法进行统计分析。两种方法都显示出$P\mathrm{中号}10$和T，尤其是在高尘季节，即从五月到九月。从动力学的角度看，相干函数突出了之间的全局相关信息$P\mathrm{中号}10$和T在频域中。为了捕获可能的局部相关信息，使用经典的基于经验模式分解（EMD）的新型互相关技术以时间尺度内在相关（TDIC）分析为基础，以多尺度方式研究两个时间序列。在计算了EMD和集成经验模式分解（EEMD）方法后，在EMD帧中检测到模式混合问题。因此，在空气污染领域首次将EEMD和TDIC结合使用。取得的结果表明，两者之间的联系的强度和性质$P\mathrm{中号}10$和T随时间尺度和时间咒语而变化。我们发现$P\mathrm{中号}10$无论是大周期性还是小周期性，来自非洲和海洋气溶胶的碳分别与T密切相关。一个人还观察到极端天气事件，例如大雨，飓风或干旱会引起强烈的反相关。