We present results of different approaches to model the evolution of the COVID-19 epidemic in Argentina, with a special focus on the megacity conformed by the city of Buenos Aires and its metropolitan area, including a total of 41 districts with over 13 million inhabitants. We first highlight the relevance of interpreting the early stage of the epidemic in light of incoming infectious travelers from abroad. Next, we critically evaluate certain proposed solutions to contain the epidemic based on instantaneous modifications of the reproductive number. Finally, we build increasingly complex and realistic models, ranging from simple homogeneous models used to estimate local reproduction numbers, to fully coupled inhomogeneous (deterministic or stochastic) models incorporating mobility estimates from cell phone location data. The models are capable of producing forecasts highly consistent with the official number of cases with minimal parameter fitting and fine-tuning. We discuss the strengths and limitations of the proposed models, focusing on the validity of different necessary first approximations, and caution future modeling efforts to exercise great care in the interpretation of long-term forecasts, and in the adoption of non-pharmaceutical interventions backed by numerical simulations.

我们展示了不同方法的结果，以模拟阿根廷 COVID-19 流行病的演变，特别关注由布宜诺斯艾利斯市及其大都市区组成的特大城市，包括总共 41 个地区，人口超过 1300 万。我们首先强调根据来自国外的感染性旅行者入境来解释流行病早期阶段的相关性。接下来，我们根据繁殖数的瞬时修改来严格评估某些提议的遏制流行病的解决方案。最后，我们构建了越来越复杂和现实的模型，从用于估计本地繁殖数量的简单同质模型，到结合手机位置数据的移动性估计的完全耦合的非齐次（确定性或随机）模型。这些模型能够通过最小的参数拟合和微调来生成与官方病例数高度一致的预测。我们讨论了所提出模型的优点和局限性，重点关注不同必要的第一近似值的有效性，并告诫未来的建模工作在解释长期预测以及采用由以下因素支持的非药物干预措施时要格外小心：数值模拟。