A novel algorithm for detecting Lake Victoria's lake-breeze fronts from station observations

Due to its vast size, Lake Victoria in East Africa significantly impacts the region’s climate through lake breezes, which influence convective activity, and thus precipitation, as well as air pollution and quality in coastal cities. These breezes, with their leading edges known as lake-breeze fronts (LBF), affect local and regional weather by initiating deep moist convection, even during the dry months of December-February (DJF) and June-August (JJA). This can lead to heavy precipitation inland or over the lake, resulting in weather-related disasters and losses. Despite their importance, there is limited research on LBFs linked with Lake Victoria. To address this gap, we developed a novel algorithm to detect LBF passages, mainly focusing on the shorelines of Uganda. We leveraged 15-minute observations from 44 automatic weather stations from the Trans-African Hydrometeorological Observatory (TAHMO) and Uganda National Meteorological Authority (UNMA) over six years (2017-2022).

Our objective observation-based lake breeze detection algorithm (OLBDA) identifies LBF passages using wind speed and direction, temperature, dew point, and precipitation measurements from stations. We focused on daytime periods (0900 – 1900LT) when the coastal land-lake temperature contrast is strongest, specifically during the dry months. The algorithm employs three criteria to detect LBF passages at any given station. First, OLBDA checks for a rapid wind reversal from offshore (relative to the nearest coastline) to onshore, or a rapid increase in wind speed within defined onshore directions. If the wind criterion is met, the data are then tested for a drop in air temperature and an increase in dew point. Here, percentile-based thresholds for temperature and dew point criteria are applied to account for regional variabilities. Lastly, to avoid false detections caused by precipitation-induced temperature drops and wind shifts, a 3-hour precipitation amount < 0.1mm at a station before the LBF passage is required. If all criteria are met, that day is considered a lake-breeze day at that station.

To test the performance, we compared the OLBDA-detected lake-breeze days with manually identified lake-breeze days (“ground truth”) within the study period from NASA’s satellite visible spectrum (Terra- and Aqua-MODIS, and NOAA-20). The algorithm detected more than 70% and 60% of the total cases identified from satellite images for coastal (within 2 km) and semi-coastal (2-10km) stations respectively, indicating good performance.

Preliminary results show that most LBF passages occur from afternoon to late evening, peaking at 1300LT for coastal stations and shifting with the station's distance from the coastline. The majority of detected lake-breeze days occur during the DJF months, with January having the most detected days. Other lake breeze characteristics including the onset and cessation time, strength, duration, propagation speed and time, and inland penetration depth, are being examined.

Finally, we aim to develop a detailed year-round observed Lake Victoria breeze climatology over Uganda. Our findings can serve as an observational benchmark to (a) improve understanding of this phenomenon’s impact on the local climate and communities along the northern shores of Lake Victoria, and (b) validate numerical simulations of lake breezes over the region.