Page 45 - EER-2-1

Explora: Environment
and Resource Statistical analysis of climate time series

over the period 1880 – 2015. The detailed statistical study 3.2. The second applications of MAL
yielded two key findings: As the previous explanation did not seem sufficient to
1. Regardless of the sub-period of 100 consecutive be entirely conclusive, the author proposed a second
months examined among the sixteen covering the explanatory hypothesis in another study. This time, the
24
1880 – 2015 time frame, the frequencies of temperature second variable considered for interaction with the global
rise or fall were approximately the same – very close to average atmospheric temperature was the heat received
equiprobability. by the Upper Oceanic Stratum (UOS) within the 0 and
2. Taking into account, the average lengths of the chains −700 m depth range.
of ascents or descents – all <1.94 and most of them
very clearly on the sixteen chains of 100 months For this analysis, he used quarterly data on the anomalies
analyzed – it was highly improbable (probability of this heat over a period of nearly 70 years (1955 – 2022),
less than 1 in 60,000) that these fluctuations were a duration long enough for a reliable study of the proposed
governed by a binomial distribution with parameters hypothesis.
n = 100 and P = 0.5. The result obtained in a given After processing the data as outlined in Section 2.2, he
month (rise or fall) appeared to significantly increase observed a distinct Markov-1 alternating behavior, similar
the probability of reversal in the following month to that seen in atmospheric temperature data, though with
(fall or rise). Hence, the author proposed that these the key difference that these data were recorded quarterly
fluctuations are better modeled as first-order Markov rather than monthly.
chains, which he describes as “alternating,” since the In summary, as detailed in Section 3.1 of this study,
24
alternation is faster than what is expected under a the author proposed the following principle of interaction:
binomial model.
The warming of the atmosphere in a given month
All the steps described in Section 2.2 were then carefully
verified, allowing the author to conclude that these chains causes evaporation, which requires calories drawn from
the UOS, resulting in its cooling. With a certain inertia
were of Markov-1 alternating type.
of the order of a month, this cooling propagates by
MAL, therefore, was already largely developed and conduction or convection into the ambient atmosphere,
finalized in this first article. which increases the probability of atmospheric cooling
The hypothesis put forward to explain this accelerated in the following month. On the contrary, the cooling of
alternation was the following: the atmosphere in a given month reduces evaporation,
1. “When the atmosphere of the globe warms up, meaning fewer calories are drawn from the UOS. As a
whatever the reason, the evaporation that this causes result, incoming heat, particularly from solar rays, is more
on the oceans and the land ends up increasing the readily retained in the UOS, causing it to warm. This heat
low cloud cover. This development of low clouds is then transferred to the surrounding atmosphere, thereby
then refreshes the atmosphere through the strong increasing the probability of atmospheric warming in the
following month.
albedo effect of these clouds, as well as through the
precipitation they bring. The quarterly, rather than monthly, alternation of
2. If, on the contrary, the globe’s atmosphere cools, there the UOS heat anomalies is explained by the author in
is less evaporation, therefore, fewer low clouds. The the following way: the thermal inertia of the UOS is
overall albedo effect of clouds decreases, which allows significantly greater than that of the atmosphere. As a
solar heat to better penetrate to the surface of the result, a duration on the order of a quarter is required for
globe (Section IV-2 of the first study ) atmospheric events to produce visible and measurable
23
repercussions on the heat received by the UOS. However,
Its validation was based only on very partial cloudiness
data (period 1983 – 2005) obtained from slide 10 of due to the accelerated monthly alternation of atmospheric
conditions – faster than what would be expected under a
Taboada. At the time, the author did not have more binomial distribution – there is a higher likelihood that,
28
complete data on cloud cover and he was aware of the over the course of a quarter, the monthly atmospheric
fragility of his conclusions since he added this:
averages follow one of the two patterns.
“It is also regrettable that we do not have such a study • Rise, then fall, then rise, leading to a greater likelihood of
over a much longer period because this would have made positive than negative results for the quarter in question.
it possible to directly study the correlation between NOAA • Fall, then rise, then fall, resulting in a greater likelihood
data and that of low cloud cover.” (Section IV-2 of the first of a negative than positive balance for the quarter in
study ). question.
23

Volume 2 Issue 1 (2025) 4 doi: 10.36922/eer.6109