Simulation models from the early COVID-19 pandemic highlighted the urgency of applying non-pharmacetical interventions (NPIs), but had limited empirical data to use. Here we use data from 2020-2021 to retrospectively model the impact of NPIs. Our model represents age groups and census division in Ontario, Canada, and is parameterised with epidemiological, testing, demographic, travel, and mobility data. The model captures how individuals adopt NPIs in response to reported cases. The model predicts that school/workplace closure and individual NPI adoption together reduced the number of deaths in the best-case scenario for the case fatality rate (CFR) from 174, 411 [CI: 168, 022, 180, 644] to 3, 383 [CI: 3, 295, 3, 483] in the Spring 2020 wave. In the Fall 2020/Winter 2021 wave, the introduction of NPIs in workplaces/schools reduced the number of deaths from 17, 291 [CI: 16, 268, 18, 379] to 4, 167 [CI: 4, 117, 4, 217]. Deaths were several times higher in the worst-case scenario for the CFR. We also estimated that each additional 7 − 11 (resp. 285 − 452) individuals who choose to adhere to NPIs in the first wave prevented one additional infection (resp., death under a best-case scenario). Our results show that the adoption of NPIs prevented a public health catastrophe.

Author contributions: VAK conceived the model and conducted data analysis. VAK and LY analyzed the model. LEB, AYK, and TM conceived and conducted the field study.VAK wrote the manuscript, with substantial input from LEB, AYK, and LY. Abstract:Modification of flow regimes and habitat degradation are the strongest, most common, and often co-occurring human activities affecting riverine populations. Ongoing efforts to restore peak flow events found under pristine flow regimes could increase advection-driven dispersal for many species. In rivers with extensive habitat loss, increased advection could transport individuals from remnant populations into degraded downstream areas, causing restored flow regimes to decrease persistence of threatened species. To resolve the capacity for such 'washout' effects across imperiled taxa, we evaluate population growth in spatial matrix models of insect, fish, and mollusc taxa experiencing advective dispersal and either long-term habitat loss or temporary disturbances. As a case study to quantify advective dispersal in threatened species, we use intensive mark-recapture methods in a Rio Grande population of the federally endangered unionid mussel Texas horhshell (Popenaias popeii). Among unionids, the most threatened freshwater taxa of North America, we find high levels of annual downstream emigration (16-51%) of adult P. popeii, concomitant with strong immigration from upstream habitats. For different taxa experiencing such advective dispersal during specific life stages, our population model shows that washout effects strongly reduce population recovery under high levels of habitat loss. Averting this negative consequence of restoring hydrology requires simultaneously restoring or protecting long, contiguous stretches of suitable habitats. Across taxa in heavily impacted systems, we suggest integrating hydrodynamic studies and field surveys to detect the presence of advective dispersal and prioritize areas for habitat restoration to enhance population persistence.