Applicability of distributed thermal sensing for identifying illicit sewage connections in urban drainage networks under tropical climates

Urban water pollution remains a major challenge for sanitation management in Brazil and other tropical regions. In areas served by separate sewer systems, illicit domestic sewage connections to stormwater drainage networks represent a significant source of contamination of urban runoff and receiving water bodies. Conventional inspection techniques for identifying such contributions are often operationally complex, spatially limited, and therefore rarely applied. Distributed temperature sensing techniques have been successfully used in temperate regions to detect sewage inputs based on thermal contrasts; however, their applicability under tropical conditions remains poorly explored.

This study investigates the thermal signature of domestic sewage in a tropical urban environment and evaluates the detectability of illicit sewage discharges in stormwater systems using a simplified thermal mixing model. Sewage temperature was monitored using thermocouples connected to a data logger with 1-minute temporal resolution in a sewer interceptor located at the São Carlos School of Engineering, University of São Paulo, Brazil, in an area characterized by student housing and food service facilities. Two monitoring campaigns were conducted. Mean sewage temperatures of 27.45 ± 0.45 °C (November 2024–April 2025) and 24.21 ± 0.54 °C (September–November 2025) were observed. A moderate Pearson correlation between sewage temperature and local air temperature (r = 0.58, p < 0.05, n = 140) indicates that atmospheric conditions partially influence sewage thermal variability.

Based on the monitored sewage temperatures (T₂) and stormwater temperature data (T₁) from the literature, a preliminary theoretical model was developed using an instantaneous energy balance approach. The model relates the detectable temperature variation (ΔT) to the sewage fraction (f), defined as the ratio between sewage discharge (Q₂) and stormwater flow (Q₁). Results indicate an exponential relationship between f and ΔT for different thermal contrasts (T₂ − T₁). The minimum detectable sewage discharge was found to be highly sensitive to ΔT, associated with the thermal resolution of the sensing system, while showing direct proportionality to stormwater flow and inverse proportionality to the thermal contrast between sewage and runoff. Future work will focus on model validation under field conditions and its extension to non-stationary flow regimes.