Characterization of Natural Regions in Guinea Using ATOVS Brightness Temperature Data Multi-Channel Microwave Analysis at 19, 23, 37, 85 and 92 GHz
Year: 2026
Type: preprint
Abstract: Passive microwave remote sensing provides unique capabilities for characterizing land surface and atmospheric conditions over tropical West Africa, where in situ observation networks remain sparse and heterogeneous. This study presents a comprehensive characterization of Guinea's four natural regions Basse Guinée (Maritime), Moyenne Guinée (Fouta Djallon), Haute Guinée (Savanna), and Guinée Forestière using multi-channel ATOVS (Advanced TIROS Operational Vertical Sounder) and MHS (Microwave Humidity Sounder) brightness temperature (T_B) data at 19, 23, 37, 85, and 92 GHz, spanning the period 2002-2022. Statistical analyses of T_B distributions reveal highly significant inter-regional differences at all channels (Kruskal-Wallis H > 796, p < 0.001; F-ANOVA > 424, p < 0.001), validating the discriminatory power of ATOVS observations. Haute Guinée consistently exhibits the warmest T_B values (T_B85 = 249.7 ± 24.6 K) while Guinée Forestière and Moyenne Guinée display the lowest brightness temperatures at high frequencies (T_B85 = 207.4 ± 31.7 K and T_B92 = 199.1 ± 33.2 K, respectively), attributable to intense convective activity and deep ice scattering over these regions. The Scattering Index (SI = T_B19 -T_B85) exhibits a strong seasonal cycle, peaking during the West African Monsoon (JJA), with maximum values observed over Guinée Forestière (SI_JJA = 38.5 ± 12.3 K). Principal Component Analysis (PCA) reveals that the first two components explain 87.4% of total variance, with PC1 representing the overall emissionscattering contrast and PC2 the vapour-emission regime. Unsupervised K-Means clustering (k = 4) recovers the four natural regions with an overall accuracy of 78.3%, demonstrating the physical relevance of the spectral space for regional classification. Long-term trend analysis (2002-2022) reveals statistically significant decreasing trends in T_B85 over Guinée Forestière (-0.30 K yr⁻ ¹; r = -0.72, p < 0.01), consistent with intensification of convective precipitation linked to global warming. These results establish ATOVS multi-channel data as a robust, operational framework for monitoring Guinea's natural regions and their response to climate variability.
Source: HAL
Institution: Université Gaston Berger
Language: English
Subfield: Environmental Engineering
Field: Environmental Science
Domain: Physical Sciences
SDG: Climate action
Open Access status: green
OpenAlex ID: W7221487232