Page 52 - EER-2-1

Explora: Environment
and Resource Statistical analysis of climate time series

the occurrence of these events. This likely explains why, for the studied climatic data. When fully implemented,
both layers, we observe a fairly balanced mix of monthly the signals obtained provide valuable information,
Markov-0 binomial type signals and Markov-1 lengthening enabling precise identification of the mechanisms and
type signals. interactions at play.
Thus, the fact that the signal may vary depending on While this method does not reveal everything or decode
the period considered (month, quarter, half-year, year) not all the climatic parameters at stake, MAL undoubtedly
only avoids real contradictions but also is well-explained serves as a valuable additional tool for enhancing our
and, in fact, contributes to a better understanding of the understanding of the climate and achieving more reliable
mechanisms at play. long-term projections.

4.3. The main reason for the effectiveness of MAL The examples described in this clearly demonstrate
the value of the method, yielding results of significant
The effectiveness of MAL lies primarily in its comparison importance in climatology, such as:
of two “spectra”—the spectra of 0 and 1 for each of the two • Detecting and explaining a phenomenon of “natural
series being analyzed. Unlike other techniques, which rely nervousness,” where global average temperature
on raw initial data or, in some cases, derived data (e.g., differences have a greater tendency to reverse their
differences of successive terms for deseasonalization), sign than maintain it from 1 month to the next. 24
MAL focuses solely on the binary aspect of whether the • Demonstrating, in what seems a definitive way, that
data increase or decrease. This transition from raw data cloudiness provides clearly negative feedback on
to their “spectrum” leads to a huge loss of information, global warming, with the OC acting as a natural
retaining only this critical piece: the direction of change. thermostat of the climate. 25
However, by isolating this single piece of information, MAL • Establishing that the parameters most critical to
prevents it from being overshadowed by other elements in defining the global climate exhibit limited long-term
the raw data, enabling it to detect signals and interactions growth, predicting temperature increases by 2100
that other statistical techniques do not detect. significantly lower than current IPCC projections. 26

4.4. Possible use of MAL in sectors other than Moreover, nothing a priori prevents MAL from being
climatology applied in other sectors beyond climatology, as long as the
system is complex and time series data are available.
Time series have long concerned many sectors other than
climatology, including econometrics, information theory, Acknowledgments
demography, astronomy, and epidemiology. 2-5,46-49
None.
Similar to its application in climatology, MAL can be
effective in other fields, provided the digital data studied Funding
meet the necessary conditions for MAL processing, None.
particularly that their successive deviations have the same
probability of being positive or negative. Additionally, the Conflict of interest
field must be complex enough to exhibit multiple and varied
interactions, a scenario where the method is well-suited The author declares no conflicts of interest.
for identifying potential interactions, as demonstrated in Author contributions
climatology. However, these conditions are restrictive, and
only practical experiments can determine whether MAL This is a single-authored article.
proves valuable outside climatology.
Ethics approval and consent to participate
It is worth noting that the concept of using Markov
chains to analyze time series data is not new; for instance, Not applicable.
it was employed as early as 1966 by Lortet-Zuckermann to Consent for publication
3
analyze a series of 444 successive explosions of the star SS
Cygni observed from 1896 to 1957. Not applicable.

5. Conclusion Availability of data
MAL is a new method for analyzing climatic time Data are available from the corresponding author upon
series based on the length of rise or fall chains in reasonable request.

Volume 2 Issue 1 (2025) 11 doi: 10.36922/eer.6109