| Journal of Collective Sciences and Sustainability
Received: 25 September 2026; Revised: 02 November 2026; Accepted: 06 November 2026; Published Online: 12 November 2026.
J. Collect. Sci. Sustain., 2026, 2(4), 26416 | Volume 2 Issue 4 (December 2026) | DOI: https://doi.org/10.64189/css.26416
© The Author(s) 2026
This article is licensed under Creative Commons Attribution NonCommercial 4.0 International (CC-BY-NC 4.0).
Analysis of Surface Meteorological Parameters
Associated with Wet Days in Maiduguri, Northeastern
Nigeria
Emmanuel Yohanna,
1,*
Gillian Musah Edward,
2
Bala Gabriel Abraham,
3
Umar Jidda Fatima
1
and Hauwa
Mohammed
4
1
Department of Physics, Kashim Ibrahim University, Maiduguri, Borno State, 602104, Nigeria
2
Department of Human Kinetic and Health Education, Taraba State University, Jalingo, Taraba State, 660213, Nigeria
3
Department of Physics, Federal University Gashua, Yobe State, 631001, Nigeria
4
Nursing Department, College of Nursing Maiduguri, Borno State, 600211, Nigeria
*Email: wabuemma@gmail.com (Emmanuel Yohanna)
Abstract
Rainfall in the Sudano-Sahelian zone of northeastern Nigeria is confined to a short monsoon season, yet daily
relationships between surface meteorological conditions and rainfall occurrence are rarely documented for
Maiduguri. This exploratory two-year (2023–2024) diagnostic study examined surface conditions associated
with wet days in Maiduguri using daily NASA POWER (MERRA-2-based) estimates of 2-m air temperature, 2-m
relative humidity, 10-m wind speed and precipitation for 2023–2024 with a number of samples (n = 731). A wet
day was defined as a day with a daily rainfall exceeding 1 mm, as rainfall alone cannot confirm thunder or
lightning, whereas wet days were not treated as thunderstorm events. The methods used included descriptive
statistics, Spearman and point-biserial correlations adjusted for serial dependence, correlations of
depersonalized anomalies, and logistic regression with seasonal harmonics and cross-year validation. In June–
September, 94.8% of the total rainfall and 89.4% of the 142 wet days contribute, and the number of wet days
increases from 56 in 2023 to 86 in 2024. Relative humidity showed the strongest association with wet-day
occurrence (r
pb
= 0.65; 95% CI: 0.57–0.72), which weakened but persisted after the seasonal cycle was removed
(r = 0.28; 95% CI: 0.17–0.38). The negative wind-speed association (r
pb
= -0.31) was no longer evident after
depersonalization, indicating a shared seasonal pattern instead of a direct physical relationship. Moreover,
surface variables provided only marginal improvements in out-of-sample discrimination beyond seasonality.
The findings describe associations, not causal effects or a validated warning tool, and justify further work
combining station, lightning and upper-air observations.
Keywords: Maiduguri; NASA POWER; Relative humidity; Sudano-Sahel; West African Monsoon; Wet days.
1. Introduction
Rainfall across the Sudano-Sahelian belt of West Africa falls within a single, short wet season and is dominated
by organized convection, particularly mesoscale convective systems and squall lines that supply most of the
seasonal total.
[1,2]
These systems also bring intense short-duration rainfall, strong outflow winds and frequent
lightning, as Africa contains several of the world’s most active lightning regions, and extreme Sahelian storms
have become more frequent since the early 1980s.
[3,4]
In northeastern Nigeria, these hazards affect rain-fed
livelihoods, transport and poorly drained urban areas.
[5,6]
Accordingly, deep convection requires low-level
moisture, conditional instability and a lifting mechanism
.[7]
Surface heating destabilizes the boundary layer, and
latent heat released during condensation sustains ascent.
[8,9]
Similarly, thunder and lightning occur when
vigorous updrafts in mixed-phase cumulonimbus clouds separate charge through collisions between graupel
and ice crystals
.[10-12]
Daily surface variables therefore describe the near-surface environment of convection but
not instability aloft or electrical activity, and a daily rainfall total cannot distinguish a thunderstorm from a
stratiform or light monsoon shower. Meteorological observations remain sparse across northern Nigeria, and
access to daily station records is often limited by cost and administrative procedures; thus, gridded satellite and
reanalysis products are widely used in the Sahel region
.[13-15]
The NASA Prediction of Worldwide Energy
Resources (POWER) service distributes continuous daily variables derived largely from the MERRA-2
reanalysis, but evaluations against Nigerian gauges show that its rainfall reproduces the seasonal cycle more
reliably than the day-to-day or interannual variability does
.[13,16-18]
Furthermore, rainfall studies in northern
Nigeria and the wider Sahel have concentrated on long-term totals, trends, onset and the annual migration of
the Intertropical Discontinuity (ITD), and recent work on Maiduguri has addressed monthly rainfall forecasting.
[19-25]
However, less attention has been given to how daily surface conditions covary with rainfall occurrence at
the city scale and to whether such associations survive once the shared seasonal cycle and the serial
dependence of daily data are considered. This is important because all four variables follow the monsoon cycle;
thus, strong correlations can arise even when day-to-day associations are weak. This study therefore provides
a short-term, exploratory diagnostic for Maiduguri based on NASA POWER data for 2023–2024. Its objectives
were to characterize the distribution and seasonal cycle of air temperature, relative humidity, wind speed,
rainfall and wet-day occurrence in each year; to quantify the associations between these variables and rainfall
occurrence; to separate seasonal from within-season coverability; and to assess whether the surface variables
add information on wet-day occurrence beyond seasonal timing. The novelty lies in this daily scale, seasonality-
adjusted treatment, in which associations are evaluated after the shared seasonal cycle is removed, with
allowance for serial dependence and cross-year validation, which are steps not usually taken in local descriptive
rainfall studies of northeastern Nigeria. Given that only rainfall was available, the outcome is defined as rainfall-
defined wet days rather than thunderstorms, and no long-term trends are inferred.
2. Study area
Maiduguri (11.85°N, 13.16°E, approximately 350 m above sea level), the capital of Borno state, lies in the Lake
Chad Basin near the borders with Niger, Chad and Cameroon, as shown in Fig. 1.
[16]
The city has a hot semiarid
climate (Köppen BSh), with a single rainy season, normally June–September, which is controlled by the
northward advance and retreat of the ITD within the West African Monsoon (WAM).
[26-28]
The rest of the year is
dominated by dry north-easterly Harmattan flow. The terrain is flat, with sparse vegetation and seasonal
drainage channels.
Fig. 1: Study area sowing (a) Nigeria and Borno State in West Africa, (b) Borno State, showing Maiduguri and the
MERRA-2 grid cell (0.5° × 0.625°) containing the NASA POWER extraction point.
3. Materials and methods
3.1 Data source and processing
Daily data for 1 January 2023 to 31 December 2024 were retrieved on 7 September 2026 from the NASA POWER
Data Access Viewer (https://power.larc.nasa.gov) for the point 11.85°N, 13.16°E.
[29]
The parameters were T2M
(air temperature at 2 m, °C), RH2M (relative humidity at 2 m, %), WS10M (wind speed at 10 m, m s
−1
) and
PRECTOTCORR (bias-corrected precipitation, mm day
−1
). All four variables are derived from MERRA-2, whose
native grid is 0.5° × 0.625°, and the precipitation is from gauge- and satellite-based observations over land.
[17,30]
The values are therefore grid-cell averages shown in Fig. 1b from the extraction point and not station
measurements. The 731-day record contained no missing dates, duplicates or physically implausible values,
and no gap filling or bias correction was applied. Python 3 (pandas, SciPy, stats models and scikit-learn) was
used for the analysis.
3.2 Representativeness of the NASA POWER data
NiMet (Nigerian Meteorological Agency) records for Maiduguri were not freely available for 2023–2024;
therefore, data quality was assessed from published evaluations. At the NiMet Maiduguri station, POWER
monthly rainfall for 1981–2015 correlated moderately with gauge records (r = 0.59; RMSE = 50.1 mm month
−1
),
whereas annual totals were essentially uncorrelated (r = 0.02), as reported by Dawa et al. (2024).
[16]
Moreover,
a study by Ellis & Samson (2026) revealed that across 20 NiMet stations, POWER reproduced spatial and
seasonal rainfall patterns but overestimated the frequency of light-rain days and underestimated wet-day
intensity.
[9]
POWER air temperatures coincide with NiMet observations in northern Nigeria, and global station
comparisons reveal modest daily skill for precipitation (NASA POWER Project, n.d.).
[31]
The data are therefore
most dependable for the seasonal cycle and relative contrasts and least dependable for exact daily amounts and
interannual totals.
3.3 Wet days
A wet day was defined as a day with rainfall exceeding 1 mm, which is consistent with the 1-mm wet-day
convention of climate indices.
[32]
The threshold excludes the trace amounts that reanalysis products generate
on many otherwise dry days, whereas in this record, 165 days had more than 0 mm but no more than 1 mm.
[13]
Wet-season rainfall in the Sudano-Sahel Basin is mainly convective, but the threshold also captures stratiform
and light rain and has no information on thunder or lightning; thus, wet days are treated only as rainfall-
occurrence days.
[2,33]
Sensitivity was tested with thresholds of 0.1, 2.5 and 5 mm and with ≥ 1 mm.
3.4 Statistical analysis
Descriptive statistics included the mean, sample standard deviation (n − 1 denominator), coefficient of variation
(CV = SD/mean × 100), quartiles, extremes and skewness, and normality was checked with the Shapiro‒Wilk
test. The monthly means, rainfall totals and wet-day counts were computed for each year, and the seasonal
shares were calculated separately for the rainfall amount and wet-day frequency. As daily rainfall was strongly
skewed and zero-inflated, associations among continuous variables were measured with Spearman’s rank
correlation (ρ), with Pearson’s r reported for comparison. The point-biserial coefficient (rpb) is used to
determine the associations with the binary wet-day series.
[34]
Daily meteorological series are serially correlated,
which inflates nominal significance, while confidence intervals (Fisher z) and p values are therefore based on
an effective sample size:
[35]
(1)
where ρ
1,x
and ρ
1,y
are the lag-1 autocorrelations of the two series. To separate seasonal from within-season
coverability, each variable was deseasonalized by subtracting a three-harmonic fit to the day of the year, and
the correlations were recomputed on the anomalies. Wet-day occurrence was modeled with logistic
regression.
[36]
Model M1 contained only two seasonal harmonics; M2 included same-day temperature, relative
humidity (per 10 percentage points) and wind speed; and M3 used previous-day values of the same variables,
as same-day means are partly altered by rain. Standard errors were autocorrelated with the Newey‒West
estimator and a 7-day lag.
[37]
The models were compared using likelihood ratio tests, the Akaike information
criterion (AIC) and the area under the receiver operating characteristic curve (AUC), and out-of-sample skill
was assessed by fitting each model to one year and testing it on the other. Statistical significance was set at p <
0.05.
4. Results and discussion
4.1 Distribution of surface meteorological variables
Table 1 summarizes the 731 daily values. Rainfall was the most variable and least normally distributed
parameter (CV = 306%, skewness = 5.22); the median was 0.00 mm, only 142 days (19.4%) were wet, and the
24 days with more than 10 mm supplied half of the two-year total. Among the remaining variables, relative
humidity had the greatest relative variability (CV = 58.0%), reflecting the contrast between Harmattan and
monsoon air, whereas temperature varied the least (CV = 14.9%). Shapiro‒Wilk tests rejected normality for all
four variables (p < 0.001), supporting the use of rank-based and model-based methods. This episodic regime is
typical of the Sudano–Sahelian zone.
[24,38]
However, because NASA POWER tends to underestimate heavy daily
totals, the maximum of 37.75 mm is seen as probably a conservative estimate of local extremes.
[13,16]
4.2 Seasonal cycle and interannual contrast
The seasonal cycle presented in Table 2 and illustrated in Fig. 2, Fig. 3 and Fig. 4 follows the WAM. June–
September supplied 878.9 mm of the 926.6 mm two-year total of 94.8% and 127 of the 142 wet days with
89.4%, two percentages that measure different quantities. August was the wettest month in both years, with 43
wet days over the two years (69.4% of August days), followed by July, with 34 days. However, the temperature
peaks in May (two-year mean of 35.4 °C), before regular rainfall begins, and then decreases as the cloud cover
and evaporative cooling increase.
[23]
However, the two-year mean relative humidity increased from
approximately 13% in March to 73% in August, whereas the mean wind speed decreased to a September
minimum and strengthened again with the return of Harmattan.
Table 1: Descriptive statistics of daily surface variables in Maiduguri, 2023–2024 (n = 731)
Statistic
Air temperature (°C)
Relative humidity (%)
Wind speed (ms
−1
)
Rainfall (mm)
Mean
29.10
38.19
2.67
1.27
Standard deviation
4.33
22.15
0.89
3.88
Coefficient of variation (%)
14.9
58.0
33.4
306.5
Minimum
18.81
6.87
0.45
0.00
25th percentile
25.96
18.29
1.99
0.00
Median
29.06
34.39
2.68
0.00
75th percentile
32.47
53.21
3.29
0.53
Maximum
38.49
90.28
5.07
37.75
Skewness
−0.09
0.53
0.19
5.22
**Note: SD is the sample standard deviation (n − 1 denominator); CV = SD/mean × 100. Skewness is the adjusted Fisher–
Pearson coefficient.
Table 2: Monthly mean temperature, relative humidity and wind speed, monthly rainfall totals and wet days (> 1
mm) in Maiduguri, 2023 and 2024
Month
Air temperature (°C)
Relative humidity (%)
Wind speed (m s
−1
)
Rainfall (mm)
Wet days
2023
2024
2023
2024
2023
2024
2023
2024
2023
2024
January
24.2
22.9
22.1
17.7
3.10
3.10
0.0
0.0
0
0
February
26.0
26.6
16.8
17.0
3.73
3.41
0.0
0.2
0
0
March
30.6
33.0
14.7
11.9
3.31
3.50
1.4
0.0
1
0
April
33.9
35.0
19.9
12.8
2.86
2.59
0.9
0.1
0
0
May
34.5
36.4
34.6
25.5
2.81
2.50
10.3
3.0
1
1
June
31.5
33.7
51.6
43.9
2.26
2.72
63.1
44.0
11
10
July
31.9
29.1
50.0
65.7
2.30
2.33
39.4
194.8
10
24
August
28.9
26.2
65.0
81.4
1.74
2.60
162.1
227.4
18
25
September
29.5
27.1
62.3
77.0
1.81
1.34
56.1
92.0
12
17
October
30.3
27.5
40.7
60.2
2.04
1.86
9.8
21.9
3
9
November
29.5
23.5
28.0
39.9
2.94
3.09
0.1
0.0
0
0
December
24.6
21.9
22.4
32.9
3.07
3.16
0.0
0.0
0
0
Annual
29.6
28.6
35.8
40.6
2.66
2.68
343.1
583.5
56
86
Jun–Sep share
(%)
–
–
–
–
–
–
93.4
95.7
91.1
88.4
**Note: The annual values are annual means (temperature, humidity, and wind speed) or annual totals (rainfall and wet
days). The two-year totals are 926.6 mm and 142 wet days, of which June–September account for 94.8% and 89.4%,
respectively.
This sequence is consistent with the northward migration of the ITD, which brings deep southwesterly
monsoon moisture to the far north of Nigeria only in mid-summer and reaches its northernmost position in
approximately August.
[22,27]
Within this moist layer, rainfall is delivered mainly by organized systems that are
often modulated by African easterly waves.
[2,39]
Hence, the rainfall frequency and peak amount are greater in
August than in April–May. The two years differed markedly, as 2024 had 86 wet days and 583.5 mm of rainfall,
compared with 56 wet days and 343.1 mm in 2023, mostly because of a wetter July–October, as presented in
Table 2 and as shown in Figs. 3 and Fig. 4. However, with two years, this contrast reflects interannual variability,
which is large in the Sahel, which does not reflect a trend.
[19,21]
The totals are of the same order as gauge evidence
for far northeastern Nigeria, where the mean wet-season rainfall is less than 500 mm and rain occurs
approximately 35 days per season.
[13]
The higher POWER wet-day counts for June–September (51 in 2023 and
76 in 2024) are consistent with the tendency of POWER to overestimate rainfall frequency even after the 1-mm
threshold is applied considering the size of the interannual difference.
[16]
Fig. 2: Monthly (a) total rainfall and mean (b) air temperature, (c) relative humidity and (d) wind speed for 2023
and 2024. The shading indicates June–September.
Fig. 3: Monthly wet days (rainfall > 1 mm) in 2023 and 2024 (bars) and two-year wet-day frequency as a percentage
of days (line). The shading indicates June–September.
Fig. 4: Daily (a) rainfall, with wet days (> 1 mm) highlighted, (b) relative humidity and (c) wind speed, 2023–2024.
The shading indicates June–September.
4.3 Associations between surface variables and rainfall occurrence
The correlations among the continuous variables where relative humidity was positively associated with daily
rainfall are shown in Table 3, and the correlation (ρ = 0.80) was much stronger than the Pearson coefficient (r
= 0.49), reflecting the skewed, zero-inflated rainfall distribution. The temperature‒humidity correlation was
n
eff
= 21 (p = 0.55). After deseasonalization, the humidity‒rainfall association decreased to ρ = 0.29 but
remained significant, whereas the wind‒rainfall correlation became weakly positive (ρ = 0.09) and the
temperature‒rainfall correlation became negative (ρ = −0.23).
Table 3: Correlations among daily surface variables before and after removal of the seasonal cycle (n = 731)
Variable pair
Raw daily values
Deseasonalized anomalies
Pearson r
Spearman ρ [95% CI]
n
eff
p
Spearman ρ [95% CI]
p
Temperature-relative humidity
−0.12
−0.14 (−0.54, 0.31)
21
0.551
−0.59 (−0.70, −0.46)
< 0.001
Temperature-wind speed
−0.17
−0.17 (−0.34, 0.01)
115
0.065
−0.10 (−0.23, 0.02)
0.112
Temperature-rainfall
−0.08
0.11 (0.00, 0.21)
344
0.040
−0.23 (−0.30, −0.15)
< 0.001
Relative humidity-wind speed
−0.50
−0.48 (−0.61, −0.33)
115
< 0.001
0.13 (0.01, 0.25)
0.037
Relative humidity-rainfall
0.49
0.80 (0.76, 0.84)
343
< 0.001
0.29 (0.22, 0.37)
< 0.001
Wind speed-rainfall
−0.19
−0.45 (−0.52, −0.37)
413
< 0.001
0.09 (0.01, 0.17)
0.026
**Note: Confidence intervals (Fisher z) and p values use the effective sample size n
eff
as in Eq. 1. Anomalies depart from
a three-harmonic seasonal fit; n
eff
for anomalies ranged from 124 to 603.
Table 4 presents associations with wet-day occurrence, where relative humidity had the strongest association
(r
pb
= 0.65, 95% CI, 0.57–0.72), wind speed had a moderate negative association (r
pb
= −0.31) and temperature
did not (r
pb
= −0.0, p = 0.47). The observed probability of a wet day was less than 3% when the humidity was
less than 40% and approximately 90% when it exceeded 80%, as illustrated in Fig. 5a; however, after the
removal of the seasonal cycle, the humidity association weakened to r = 0.28 (95% CI, 0.17–0.38) but persisted,
the temperature association became negative (r = −0.19), and the wind association was no longer evident (r =
−0.01, p = 0.90), as shown in Fig. 5b. The results were similar for thresholds of 0.1–5 mm (humidity r
pb
= 0.77–
0.44).
Table 4. Associations between surface variables and wet-day occurrence (142 wet and 589 nonwet days)
Predictor
r
pb
(95% CI)
p
Deseasonalised r [95% CI]
p
Univariate OR [95% CI]
AUC
Air temperature (per
1 °C)
−0.05 (−0.17, 0.08)
0.470
−0.19 (−0.30, −0.07)
0.002
0.97 (0.91, 1.04)
0.551
Relative humidity (per
10%)
0.65 (0.57, 0.72)
< 0.001
0.28 (0.17, 0.38)
< 0.001
2.90 (2.42, 3.48)
0.934
Wind speed (per 1 m
s
−1
)
−0.31 (−0.40, −0.21)
< 0.001
−0.01 (−0.11, 0.09)
0.900
0.36 (0.25, 0.51)
0.729
**Note: r
pb
is the point-biserial correlation, and p values and intervals use the effective sample size. The
deseasonalized coefficient correlates predictor anomalies with wet-day occurrence. OR = odds ratio from
univariate logistic regression with Newey‒West (7-day) standard errors, AUC = area under the ROC curve.
The logistic models indicate this pattern, as presented in Table 5, as the seasonal timing (M1) discriminates wet
from nonwet days (AUC = 0.910). However, adding same-day variables (M2) improved the fit significantly
(likelihood ratio χ
2
= 76.8, p < 0.001). Each 10-percentage-point increase in humidity multiplied the odds of a
wet day by 5.5 (95% CI, 3.2–9.7), whereas temperature and wind speed were less significant. Moreover,
previous-day humidity (OR = 3.6) and temperature (OR = 1.5) were significant in M3. The out-of-sample gain
was small and negligible, whereas the AUC increased from 0.932 to 0.949 when a 2023-trained model was
tested on 2024, only from 0.881 to 0.885 in the reverse direction, and M3 did not outperform M1 for 2023.
Table 5: Logistic regression models of wet-day occurrence (n = 730 days)
Model
Predictors
AUC
AUC
2023→2024
AUC
2024→2023
AIC
LR χ
2
vs M1 (p)
Odds ratio [95% CI]
M1
Seasonal
harmonics only
0.910
0.932
0.881
421.1
–
–
M2
M1 + same-day
T, RH, WS
0.945
0.949
0.885
350.4
76.8 (< 0.001)
RH 5.55 (3.16, 9.74), T 1.29 (0.96,
1.73), WS 0.82 (0.53, 1.27)
M3
M1 + previous-
day T, RH, WS
0.923
0.939
0.868
399.5
27.7 (< 0.001)
RH 3.59 (2.14, 6.02), T 1.46 (1.10,
1.95), WS 0.92 (0.61, 1.38)
**Note: T = air temperature (per 1 °C), RH = relative humidity (per 10%), WS = wind speed (per 1 m s
−1
). Seasonal
harmonics are two sine-cosine pairs of the day of the year. AUC 2023→2024 denotes a model fitted to 2023 and tested
on 2024. The likelihood ratio (LR) tests have 3 degrees of freedom. Standard errors use the Newey‒West estimator (7-
day lag); the first day is excluded, as M3 requires previous-day values.
Fig. 5: (a) Proportion of wet days by relative humidity class (95% Wilson intervals, class sizes above) with the fitted
logistic curve and 95% confidence band. (b) Standardized deseasonalized anomalies on nonwet and wet days.
Relative humidity therefore provides information on rainfall occurrence beyond the seasonal calendar, which
is consistent with the importance of low-level moisture for Sahelian convection.
[40]
Although this does not
establish that humidity controls storm formation, as same-day humidity is caused by rain and clouds and
instability, convective inhibition and wind shear, which determine whether moist air develops into deep
convection, were outside the scope of the data. Cooler anomalies on wet days are likewise more plausibly a
consequence of rain than a precursor.
[4]
The negative wind–speed association is not evidence that calm
conditions precede convection. It is not evident after deseasonalization and was absent from June–September
(r
pb
= −0.02), indicating that it reflects stronger dry-season Harmattan winds, which coincide with the absence
of rain. Moreover, daily winds averaged over a grid cell also smooth out the short-lived gust fronts and squall-
line outflows that accompany Sahelian storms.
[5,33]
5. Implications for seasonal risk communication
With respect to the timing of rainfall, approximately nine of the ten wet days fell between June and September,
peaking in August, while seasonal preparedness in Maiduguri, such as reminders on drainage maintenance,
flood readiness and lightning safety, would therefore be best issued before June and reinforced in July–August,
alongside the predictions and warnings of NiMet.
[41]
However, lightning and public awareness were not
observed in the data used, and the humidity results were not proposed as warning criteria.
6. Limitations and future research
The exploratory scope of this study points to clear directions for further research. The two-year record captures
the seasonal cycle and one marked interannual contrast. Extending the analysis to the full NASA POWER record
from 1981 onward would place 2023–2024 in a climatological context. Concurrent daily NiMet observations
and upper-air indices, such as convective available potential energy, convective inhibition, lapse rates and
vertical wind shear, were not available during this study. Comparing POWER with station records and adding
ERA5 or MERRA-2 convective indices would strengthen data validation and the physical interpretation of the
humidity association.
[7,40]
7. Conclusion
In this study, two years of NASA POWER data were used to examine surface meteorological conditions
associated with rainfall-defined wet days in Maiduguri. The rainfall was concentrated during the monsoon
season; specifically, 94.8% of the rainfall and 89.4% of the wet days were supplied during June–September,
with a peak in August, and the greater number of wet days in 2024 reflects the interannual variability. Relative
humidity was the surface variable most closely associated with wet-day occurrence, and part of this association
persisted after the seasonal cycle was removed. The negative association with wind speed was entirely
seasonal, and temperature showed no consistent relationship, while surface variables added modest
information beyond seasonal timing. Furthermore, by separating seasonal associations from within-season
associations, this study provides a transparent local baseline for Maiduguri and a reproducible framework that
can be extended with longer records, station data, convective indices and lightning observations. Future studies
could incorporate satellite or ground-based lightning observations and radar to allow thunderstorm days to be
distinguished from other rainy days.
Acknowledgments
The study data were obtained from the NASA Langley Research Center POWER Project funded through the
NASA Earth Science Directorate Applied Science Program.
CRediT Author Contribution Statement
Emmanuel Yohanna: Conceptualization, Formal analysis, Writing – Review & editing. Gillian Musah Edward:
Formal analysis, Writing – Review & editing. Bala Gabriel Abraham: Conceptualization, Formal analysis,
Supervision, Writing – Review & editing. Umar Jidda Fatima: Formal analysis, Writing – Review & editing.
Hauwa Mohammed: Formal analysis, Writing – Review & editing. All authors have read and approved the final
version of the manuscript for publication and agree to be accountable for all aspects of the work, ensuring that
questions related to the accuracy or integrity of any part of the work are appropriately investigated and
resolved.
Funding Declaration
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-
profit sectors.
Data Availability Statement
The daily data are freely available from NASA POWER (https://power.larc.nasa.gov) using the parameters and
coordinates given in Section 3.1.
Competing Interests
There is no conflict of interest.
Artificial Intelligence (AI) Use Disclosure
The authors declare that artificial intelligence (AI)-assisted tools were used only for language refinement,
grammar improvement, and manuscript structuring purposes during the preparation of this work. All technical
content, experimental implementation, results, and interpretations were independently developed and verified
by the authors.
Supporting Information
Not applicable.
References
[1]
A. H. Fink, D. G. Vincent, V. Ermert, Rainfall types in the West African Sudanian zone during the summer
monsoon 2002, Monthly Weather Review, 2006, 134, 2143–2164, doi: 10.1175/MWR3182.1.
[2]
V. Mathon, H. Laurent, T. Lebel, Mesoscale convective system rainfall in the Sahel, Journal of Applied
Meteorology, 2002, 41, 1081–1092, doi: 10.1175/1520-0450(2002)0412.0.CO;2.
[3]
R. I. Albrecht, S. J. Goodman, D. E. Buechler, R. J. Blakeslee, H. J. Christian, Where are the lightning
hotspots on Earth?, Bulletin of the American Meteorological Society, 2016, 97, 2051–2068, doi:
10.1175/BAMS-D-14-00193.1.
[4]
C. M. Taylor, D. Belušić, F. Guichard, D. J. Parker, T. Vischel, O. Bock, P. P. Harris, S. Janicot, C. Klein, G.
Panthou, The frequency of extreme Sahelian storms has tripled since 1982 according to satellite
observations, Nature, 2017, 544, 475–478, doi: 10.1038/nature22069.
[5]
O. J. Matthew, M. A. Ayoola, Seasonality of wind speed, wind shears and precipitation over West Africa,
Journal of Atmospheric and Solar-Terrestrial Physics, 2020, 207, 105371, doi:
10.1016/j.jastp.2020.105371.
[6]
D. Gueye, A. L. Dieng, K. M. Núñez Ocasio, Z. L. Moon, A. Deme, M. Sarr, S. Diouf, A. A. Coly, P. Fall, M.
Thiam, Characterizing the morphology and thermodynamics of mesoscale convective systems driving
extreme rainfall in West Africa, in West African Conference on Renewable Energies, Springer Nature
Switzerland, 2025, 429–451, doi: 10.1007/978-3-032-17129-0_28.
[7]
C. A. Doswell III, Severe convective storms—An overview, in C. A. Doswell III (Ed.), Severe Convective
Storms, 2001, 28,1–26, doi: 10.1175/0065-9401-28.50.1.
[8]
D. M. Romps, Ascending columns, WTG, and convective aggregation, Journal of the Atmospheric
Sciences, 2021, 78, 497–508, doi: 10.1175/JAS-D-20-0041.1.
[9]
B. A. Choudhury, M. I. R. Tinmaker, Thermodynamic control of lightning activity in premonsoon and
monsoon season over the Indian region, Journal of Atmospheric and Solar-Terrestrial Physics, 2025,
268, 106410, doi: 10.1016/j.jastp.2024.106410.
[10]
D. R. MacGorman, W. D. Rust, The Electrical Nature of Storms, Oxford University Press, 1998.
[11]
V. A. Rakov, M. A. Uman, Lightning: Physics and Effects, Cambridge University Press, 2003.
[12]
C. Saunders, Charge separation mechanisms in clouds, Space Science Reviews, 2008, 137, 335–353, doi:
10.1007/s11214-008-9345-0.
[13]
A. W. Ellis, B. Samson, Comparative assessment of POWER and CMAP satellite-derived rainfall
products for hydroclimate monitoring in Nigeria, Theoretical and Applied Climatology, 2026, 157, 228,
doi: 10.1007/s00704-026-06121-1.
[14]
J. F. Hermance, H. M. Sulieman, Adequacy of the daily TMPA 3B42 high-resolution satellite
precipitation product for monitoring hydrometeorological hazards in the Southeast Sahel of Africa,
International Journal of Remote Sensing, 2018, 39, 2579–2596, doi:
10.1080/01431161.2018.1430405.
[15]
E. K. Vizy, K. H. Cook, Extreme rainfall events in the West African Sahel: Understanding storm
development over the Damergou gap using convection-permitting simulations in the Weather
Research and Forecasting model, Quarterly Journal of the Royal Meteorological Society, 2023, 149,
959–983, doi: 10.1002/qj.4443.
[16]
S. Y. Dawa, M. L. Tan, N. Samat, R. Roy, F. Zhang, Evaluation of five gridded precipitation products for
estimating precipitation and drought over Yobe, Nigeria, Water Supply, 2024, 24, 2039–2054, doi:
10.2166/ws.2024.113.
[17]
R. Gelaro, W. McCarty, M. J. Suárez, R. Todling, A. Molod, L. Takacs, C. A. Randles, A. Darmenov, M. G.
Bosilovich, R. Reichle, K. Wargan, The Modern-Era Retrospective Analysis for Research and
App lica tion s, Ve rs io n 2 ( MER RA -2 ), Jo ur n al of Cl imate , 20 17, 30, 5 419 –54 54,
doi: 10.1175/JCLI-D-16-0758.1.
[18]
J. C. Mikovitz, P. W. Stackhouse Jr., F. Patadia, B. Macpherson, N. Khadka, A Global Agroclimatology
Solar Insolation and Meteorological Parameter Data Base: Meteorological Evaluation and Application
Updates to NASA POWER, conference presentation, ASA, CSSA and SSSA International Annual Meeting,
2023.
[19]
A. Bichet, A. Diedhiou, West African Sahel has become wetter during the last 30 years, but dry spells
are shorter and more frequent, Climate Research, 2018, 75, 155–162, doi: 10.3354/cr01515.
[20]
A. Maina, M. Grema, Application of an STL–ARIMA hybrid framework for monthly rainfall forecasting
in Maiduguri, Nigeria, Journal of Science Research and Reviews, 2026, 3, 87–98, doi:
10.70882/josrar.2026.v3i1.146.
[21]
S. E. Nicholson, The West African Sahel: A review of recent studies on the rainfall regime and its
interannual variability, International Scholarly Research Notices, 2013, 2013, 453521, doi:
10.1155/2013/453521.
[22]
T. O. Odekunle, F. A. Adesina, I. E. Gbode, A. O. Adeyefa, S. O. Ajadi, Variabilities and trends in the onset,
peak, and retreat of Intertropical Discontinuity (ITD) annual migrations over West Africa between
1979 and 2020, Climate Dynamics, 2024, 62, 7149–7169, doi: 10.1007/s00382-024-07270-7.
[23]
B. F. Sasanya, S. O. Adesogan, A. A. Ademola, Spatiotemporal analysis of rainfall over Chad River Basin,
Nigeria, Theoretical and Applied Climatology, 2025, 156, 118, doi: 10.1007/s00704-024-05338-2.
[24]
A. Tarhule, M. K. Woo, Changes in rainfall characteristics in northern Nigeria, International Journal of
Climatology, 1998, 18, 1261–1271, doi: 10.1002/(SICI)1097-0088(199809)18:11<1261::AID-
JOC302>3.0.CO;2-Z.
[25]
E. Yohanna, S. U. Abubakar, F. U. Jidda, Assessing the impacts of temperature and precipitation
dynamics on climate variability across Northeast Nigeria, International Journal of Environment and
Climate Change, 2025, 15, 121–136.
[26]
D. Cui, S. Liang, D. Wang, Observed and projected changes in global climate zones based on Köppen
climate classification, WIREs Climate Change, 2021, 12, e701, doi: 10.1002/wcc.701.
[27]
M. I. Lélé, P. J. Lamb, Variability of the Intertropical Front (ITF) and rainfall over the West African
Sudan–Sahel zone, Journal of Climate, 2010, 23, 3984–4004, doi: 10.1175/2010JCLI3277.1.
[28]
B. Sultan, S. Janicot, The West African monsoon dynamics. Part II: The “preonset” and “onset” of the
summer monsoon, Journal of Climate, 2003, 16, 3407–3427, doi: 10.1175/1520-
0442(2003)0162.0.CO;2.
[29]
A. H. Sparks, nasapower: A NASA POWER global meteorology, surface solar energy and climatology
data client for R, Journal of Open Source Software, 2018, 3, 1035, doi: 10.21105/joss.01035.
[30]
R. H. Reichle, Q. Liu, R. D. Koster, C. S. Draper, S. P. P. Mahanama, G. S. Partyka, Land surface
precipitation in MERRA-2, Journal of Climate, 2017, 30, 1643–1664, doi: 10.1175/JCLI-D-16-0570.1.
[31]
O. T. Faloye, V. Kamchoom, A. T. Ogunrinde, A. E. Ajayi, P. G. Oguntunde, N. Sinsamutpadung, Evaluation
of various machine learning-based bias correction approaches for NASA POWER air temperatures: A
case study of Nigeria, Big Earth Data, 2025, 9, 1111–1135, doi: 10.1080/20964471.2025.2565884.
[32]
X. Zhang, L. Alexander, G. C. Hegerl, P. Jones, A. Klein Tank, T. C. Peterson, B. Trewin, F. W. Zwiers,
Indices for monitoring changes in extremes based on daily temperature and precipitation data, WIREs
Climate Change, 2011, 2, 851–870, doi: 10.1002/wcc.147.
[33]
NASA POWER Project, Meteorological Data Assessment Process, NASA Langley Research Center, n.d.,
https://power.larc.nasa.gov/docs/methodology/meteorology/assessment/, Accessed 20 September
2026.
[34]
D. S. Wilks, Statistical Methods in the Atmospheric Sciences, 4th ed., Elsevier, 2019, doi: 10.1016/C2017-
0-03921-6.
[35]
C. S. Bretherton, M. Widmann, V. P. Dymnikov, J. M. Wallace, I. Bladé, The effective number of spatial
degrees of freedom of a time-varying field, Journal of Climate, 1999, 12, 1990–2009, doi:
10.1175/1520-0442(1999)0122.0.CO;2.
[36]
D. W. Hosmer Jr., S. Lemeshow, R. X. Sturdivant, Applied Logistic Regression, 3rd ed., Wiley, 2013, doi:
10.1002/9781118548387.
[37]
W. K. Newey, K. D. West, A simple, positive semidefinite, heteroskedasticity and autocorrelation
consistent covariance matrix, Econometrica, 1987, 55, 703–708, doi: 10.2307/1913610.
[38]
M. V. K. Sivakumar, Agroclimatic Aspects of Rainfed Agriculture in the Sudano-Sahelian Zone, ICRISAT,
1989, http://oar.icrisat.org/id/eprint/4464, Accessed 15 September 2026.
[39]
A. H. Fink, A. Reiner, Spatiotemporal variability of the relation between African easterly waves and
West African squall lines in 1998 and 1999, Journal of Geophysical Research: Atmospheres, 2003, 108,
4332, doi: 10.1029/2002JD002816.
[40]
D. J. Parker, M. Diop-Kane, Meteorology of Tropical West Africa: The Forecasters’ Handbook, Wiley,
2017, doi: 10.1002/9781118391297.
[41]
M. A. Cooper, C. J. Andrews, R. L. Holle, R. Blumenthal, N. Navarrete-Aldana, Lightning-related injuries
and safety, in P. S. Auerbach (Ed.), Auerbach’s Wilderness Medicine, 2017, 7, 71–117.
Publisher Note: The views, statements, and data in all publications solely belong to the authors and
contributors. GR Scholastic is not responsible for any injury resulting from the ideas, methods, or products
mentioned. GR Scholastic remains neutral regarding jurisdictional claims in published maps and institutional
affiliations.
Open Access
This article is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License, which
permits the non-commercial use, sharing, adaptation, distribution and reproduction in any medium or format,
as long as appropriate credit to the original author(s) and the source is given by providing a link to the Creative
Commons License and changes need to be indicated if there are any. The images or other third-party material
in this article are included in the article's Creative Commons License, unless indicated otherwise in a credit line
to the material. If material is not included in the article's Creative Commons License and your intended use is
not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly
from the copyright holder. To view a copy of this License, visit: https://creativecommons.org/licenses/by-
nc/4.0/
© The Author(s) 2026
