Super-Resolution Bias-Correction (SRBC) for Digital Epidemiology: Downscaling Noisy Signals to Actionable Geographies

Digital signals are often coarse (state/country-level Google Trends) or demographically skewed, limiting local actionability. Climate researchers have advanced SRBC for rainfall using deep learning (Singh et al., 2023), and hydrologists correct systematic biases in remote-sensed fields (e.g., Thomé Brochado & Rennó, 2024). We propose an SRBC pipeline that learns mappings from coarse digital indicators to fine-grained incidence proxies by: (1) conditioning on local anchors (wastewater dPCR estimates modeled correctly per Lison et al., 2025; clinic line lists), (2) injecting demographics and health-care access covariates to correct selection bias, and (3) applying simulation-based bias correction (indirect inference/iterative bootstrap per Guerrier et al., 2018) to remove residual estimator bias in the downscaled fields. Compared to standard spatial interpolation or naive risk redistribution, this approach explicitly models and corrects both scale mismatch and systematic bias. The payoff is neighborhood-level, bias-adjusted digital surveillance layers that can drive targeted interventions—especially valuable where traditional surveillance is sparse.

References:

1. Real-time outlier detection in digital PCR data for wastewater-based pathogen surveillance. R. E. McLeod, T. R. Julian, T. Stadler, A. Lison (2025). medRxiv.
2. Improving inference in environmental surveillance by modeling the statistical features of digital PCR. A. Lison, Timothy R. Julian, T. Stadler (2025). bioRxiv.
3. An Intercomparison of Deep-Learning Methods for Super-Resolution Bias-Correction (SRBC) of Indian Summer Monsoon Rainfall (ISMR) Using CORDEX-SA Simulations. Deveshwar Singh, Yunsoo Choi, Rijul Dimri, Masoud Ghahremanloo, Arman Pouyaei (2023). Asia-Pacific Journal of Atmospheric Sciences.
4. New Method to Correct Vegetation Bias in a Copernicus Digital Elevation Model to Improve Flow Path Delineation. Gabriel Thomé Brochado, C. D. Rennó (2024). Remote Sensing.
5. Simulation-Based Bias Correction Methods for Complex Models. S. Guerrier, E. Dupuis-Lozeron, Yanyuan Ma, Maria-Pia Victoria-Feser (2018). Journal of the American Statistical Association.