key: cord-1022870-8nxe615b authors: Smith, T. P.; Dorigatti, I.; Mishra, S.; Volz, E.; Walker, P. G. T.; Ragonnet-Cronin, M.; Tristem, M.; Pearse, W. D. title: Environmental drivers of SARS-CoV-2 lineage B.1.1.7 transmission in England, October to December 2020 date: 2021-03-12 journal: nan DOI: 10.1101/2021.03.09.21253242 sha: 7537a02a720f1e33f9d19c859d540a1a8a48be70 doc_id: 1022870 cord_uid: 8nxe615b Previous work has shown that environment affects SARS-CoV-2 transmission, but it is unclear whether emerging strains show similar responses. Here we show that lineage B.1.1.7 spread with greater transmission in colder and more densely populated parts of England. We also find evidence of B.1.1.7's transmission advantage at warmer temperatures versus other strains, implying that spring conditions may facilitate B.1.1.7's invasion in Europe and across the Northern hemisphere, undermining the effectiveness of public health interventions. transmission is still key for the optimal implementation of control strategies. SARS-CoV-2 37 transmission is influenced by local environmental conditions [1] , and we have previously 38 shown that colder temperatures and higher population densities explain spatial and 39 temporal variation in transmission intensity [2] . A critical insight from previous work has 40 been that human behaviour drives SARS-CoV-2 transmission, and that environment only 41 plays a marginal effect when effective non-pharmaceutical interventions (NPIs) are in place 42 [1, 2] the impact of the environment by investigating the association between weekly R estimates 63 and the predictors (i.e., spatial variation in R), for sequential weeks. To account for the 64 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 12, 2021. ; https://doi.org/10.1101/2021.03.09.21253242 doi: medRxiv preprint VOC's emergence in the South-East (which is among the warmest and most densely 65 populated parts of England), and its subsequent spread into colder, less densely populated 66 parts of the country, we incorporated interactions between VOC frequency and each of 67 temperature and population density. Without including this interaction, any effect of 68 temperature or population density on R might be compounded or masked by spatial 69 autocorrelation with VOC frequency. Following Volz et al. [7], we measured the effects of 70 the weekly predictors on R in the subsequent week, to account for the generation time of 71 SARS-CoV-2. 72 As observed by Volz et al. [7], we find a strong positive effect of VOC frequency on 73 transmission intensity, particularly during the initial emerging phase when VOC frequency 74 was less consistent among regions (Table 1) . While the UK's national lockdown initially 75 decreased SARS-CoV-2 transmission intensity, as the VOC spread, R increased across 76 regions ( Figure 1 ). We find that the impact of the environment on transmission is mediated 77 by VOC frequency, and it is only at higher VOC frequencies that the effects of temperature 78 and population density are pronounced (Table 2 and Figure 1 ). Furthermore, we find a 79 greater environmental effect after the UK moved from full lockdown to a (less strict) tiered 80 system, confirming our previous observations that the effect of temperature becomes 81 pronounced when NPIs are relaxed [2]. Thus, the effects of environmental drivers, and 82 even those of the increased transmissibility of the VOC, are secondary to differences in 83 human behaviour driven by differences in NPIs. As the VOC has now spread to dominate 84 the country's viral population, we now expect environmental factors to become the 85 dominant driver of spatial variation in R, with colder areas facing higher transmission 86 intensities than warmer regions, unless NPIs or the accumulation of immunity (either 87 naturally acquired or vaccine derived) sufficiently reduce transmission. 88 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The role of natural immunity 118 To investigate the potential role of immunity in our results, we perform a sensitivity 119 analysis and re-run our regression analyses on R estimates corrected for the relative attack 120 rate (AR) estimates. This accounts for the potential slowdown in R due to the accumulation 121 of natural immunity across the population through time. We obtain qualitatively identical 122 results in the AR-corrected analyses, but generally lower correlation (additive decrease in 123 2 of 16% in weeks 45 and 46) as the VOC emerged, versus in the later stages (average 124 decrease in 2 of only 3%, Supplementary Table S1). Note that the timeframe of our 125 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 12, 2021. ; https://doi.org/10.1101/2021.03.09.21253242 doi: medRxiv preprint analyses was prior to mass vaccination of the UK public and even in the latest timepoint 126 (week 50), median attack rate across regions was less than 10% (Supplementary Figure 127 S1 ). 128 Using the model fitted to the week 50 data, our results suggest that when the variant 150 accounts for 100% of SARS-CoV-2 cases in the population, R increases by approximately 151 0.12 per ℃ of temperature decrease. However, we also find that in warmer conditions, the 152 difference in transmissibility between VOC and non-VOC strains is greater. Our results 153 suggest that for each ℃ of temperature increase, the ratio of VOC to non-VOC R may 154 increase by approximately 0.12-0.22 (Supplementary Table S2 ). 155 While we caution against extrapolating from winter to summer conditions, our results 156 highlight that there is no reason to suppose that summer weather alone will slow down the 157 invasion dynamics of B.1.1.7 and significantly reduce the transmission intensity of SARS-158 CoV-2. Thus, it is imperative to quantify and continue to monitor the impact of 159 interventions (e.g., vaccines and NPIs) to inform policy. In that regard, the speed with 160 which the VOC has spread through the UK, while concerning from a public health 161 perspective, provides the perfect opportunity to parameterise models of its responses. 162 Importantly in the context of new strains, transmission of respiratory viruses in tropical 163 climates (such as Brazil, where a new variant has arisen [4]) are often more sensitive to 164 rainfall than temperature [10] . Better characterisation of past, and new potential 165 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted March 12, 2021. ; https://doi.org/10.1101/2021.03.09.21253242 doi: medRxiv preprint relationships between environment and transmission, as new variants emerge and spread 166 internationally, is a research priority. 167 The Role of 170 Environmental Factors on Transmission Rates of the COVID-19 Outbreak: An Initial 171 173 Environment influences SARS-CoV-2 transmission in the absence of non-pharmaceutical 174 interventions Preliminary 178 genomic characterisation of an emergent SARS-CoV-2 lineage in the UK defined by a novel 179 set of spike mutations. Virological Genomic characterization of a novel SARS-CoV-2 lineage from Rio de Janeiro, Brazil. 183 medRxiv 187 Emergence and rapid spread of a new severe acute respiratory syndrome-related 188 coronavirus 2 (SARS-CoV-2) lineage with multiple spike mutations in South Africa Genomics and 193 epidemiology of a novel SARS-CoV-2 lineage in Manaus, Brazil. medRxiv [preprint Transmission of 198 SARS-CoV-2 Lineage B.1.1.7 in England: Insights from linking epidemiological and genetic 199 data A COVID-19 Model for 203 Local Authorities of the United Kingdom Interpretation of Models of Fundamental Ecological Niches and 207 Species' Distributional Areas Environmental factors affecting the transmission of respiratory 209 viruses Curr Opin Virol