Developing a long-term high-resolution winter fog climatology over south Asia 1 using satellite observations from 2002 to 2020 2 3

The vast Indo-Gangetic Plains (IGP) south of the Himalaya are subject to dense fog every year during winter months (December-January), severely disrupting rail, air and public transport of millions of people living in northern India, Pakistan, Nepal and Bangladesh. Air pollution combined with high moisture availability in the shallow boundary layer, are important factors affecting the persistence and widespread nature of fog over the IGP. Despite the environmental significance and impacts on the public at-large, an in depth understanding of the long-term spatial-temporal distribution of the south Asian fog, is presently not available in the literature. We utilize infrared remote sensing techniques to develop a high-resolution (≈1 km x 1 km) fog detection climatology over the past two decades (2002 – 2020), using Aqua/MODIS satellite observations. A dynamic brightness temperature difference threshold (involving 3.96 μm and 11.03 μm bands) for nighttime fog detection is constructed based on systematic radiative transfer simulations involving cloud effective radius, cloud top height, cloud optical depth and satellite viewing geometry. Our satellite-based fog detection is consistent with theoretical simulations of fog characterization and is also found to be well-correlated with near-surface visibility observations of dense fog (r = 0.87, p-value << 0.01). We also provide satellite-derived nighttime estimates of fog/low-cloud effective radius which is in general agreement with the operational daytime MODIS cloud data product and limited in situ observations. In terms of fog frequency, the IGP is relatively uniformly covered by widespread fog occurrences with the largest frequency found in the low-lying Terai region, bordering India and Nepal, which is also consistently observed in our daytime fog detection results over the last two decades. Additionally, the interannual variations in fog occurrences track closely with that of relative humidity in the IGP, which is associated with shallow boundary layer conditions during winter-time favoring the formation and persistence of fog. Overall, these long-term satellite-derived results present new high-resolution data and insights into the dense and often intense winter fog occurrences which routinely engulf the entire stretch of the Indo-Gangetic Plains and cause significant degradation to ground visibility in one of the world’s most densely populated regions.


39
Each year during the winter months (December-January), the Indo-Gangetic Plains (IGP)  In this study, we utilize daily high-resolution satellite observations acquired during the past two 66 decades, along with inputs from surface meteorological observations as well as model and 67 4 observation-based reanalysis datasets, to map and quantify the spatial and temporal distribution 68 of the dense fog cover over the IGP. 69 Research on fog detection using satellite remote sensing has been carried out mainly 70 using multi-spectral systems involving thermal infrared channels (Gultepe et al., 2007). Hunt

71
(1973) suggested that small droplets found in fog are associated with lower emissivity at 3.7 μm 72 than at 10.8 μm, while the emissivity at these two bands is roughly the same for larger droplets.

73
The difference in emissivity leads to significant contrast in brightness temperatures at the mid- Optical Thickness (COT) and CER, derived using visible radiances, were also used for daytime 115 fog characterization. This product include datasets at a spatial resolution of 1 km or 5 km. The

116
CER retrieval is obtained via a dual-channel retrieval method with band 7 (2.1 μm) combined  3. Nighttime fog remote sensing 134 We primarily discuss here the nighttime fog detection framework using MODIS observations 135 over the IGP. In addition, we leverage an existing approach for daytime fog detection and  In order to characterize changes in cloud droplet size and its impact on at-sensor satellite 181 radiance/brightness temperatures, the simulated values at = 20° for 3.96 and 11.02 , 182 with respect to cloud effective radius are shown in Fig. 3. The Fig. 3a  Ghude et al., 2017). The simulated ∆ corresponding to < 9 were found to be larger than 195 2.5 K (Fig. 3b). A fixed ∆ threshold is presently being used by the Indian Meteorological computations. Fig. 4a shows the variation in with (2 ≤ ≤ 40 ) and ( = 5, 10, 20, 204 30) at 3.96 and 11.03 . The at smaller is greater than the at higher at both the 205 SWIR and IR channels, due to the attenuation of the nighttime radiance as the opacity of the 206 cloud increases. Fig. 4b shows as a function of and ℎ with 2 ≤ ≤ 40 and ℎ varying 207 from 1 km to 4 km. The top of the fog layer at higher altitudes is cooler than at lower altitudes, 208 which is evident in the lower brightness temperature at ℎ = 4 km than at ℎ = 1 km, for both 209 3.96 and 11.03 . There is also a small but non-negligible variation present in ∆ 210 corresponding to ℎ for < 9 (Fig. 4e). In addition, we find that the ∆ is smaller than 2 K 211 for fog layers with higher cloud tops and low cloud optical depth (for > 9 ), further 212 indicating that low ∆ is plausible with cloud layer at high altitudes and larger . Finally, the 213 largest sensitivity is found for the satellite viewing geometry where Fig. 4c shows the variation 214 in the Tb with (0°≤ ≤ 60°). The brightness temperature at both the channels, 3.96 and 215 11.02 , especially at the shorter wavelength, drops significantly at larger , where the 216 distance between the sensor and the fog/low-cloud feature is greater than that at smaller .

217
Specifically, the emitted radiation from the surface-fog feature passes through a longer 218 atmospheric path at larger , which leads the brightness temperature to be cooler, compared to 219 lower temperature at smaller . There is a pronounced variation in ∆ with ; for example at 9 220 , the ∆ is 4.5 K at = 40°, significantly larger than ∆ of 2.5 K at = 20° (Fig. 4f).  observations. The maps were created at 0.01° x 0.01° spatial resolution (approximately 1 km x 1 247 km resolution). We also provide here an estimation of the CER for fog-detected pixels (relevant 248 information is provided in supplementary material). In Fig. 6b and 6c, the spatial distribution of where the distribution is generally found to be consistent with the daytime CER. 254 We then expanded the processing of daily nighttime fog detection to develop a long-term   We also analyzed the interannual variations in satellite-derived fog frequency with 301 ground-observed poor visibility conditions associated with fog (visibility < 250 m) averaged 302 over the 9 meteorological stations, and found a significantly high correlation of 0.93 (p-value << 303 0.01) (Fig. 9). The year-to-year variations in fog are found to be well correlated with monthly

309
In this study, we use 19 years of satellite observations to produce a high-resolution (≈1 properties of the two channels for fog droplets. We used a radiative transfer framework involving 315 15 satellite radiances and existing retrievals of cloud properties to map and quantify fog detections.

316
Here, we specifically constructed a dynamical threshold for fog detection based on brightness 317 temperature differences as a function of various satellite and fog/low-cloud parameters including 318 viewing geometry, fog effective radius, fog vertical distribution and its optical thickness. In 319 addition to fog detection, we also characterize size of fog/low-cloud droplets in terms of their 320 effective radius which is found to be less than 9 µm (mean=7.2 and standard deviation=1.1). To