**1. Introduction**

The COVID-19 outbreak occurred just before the 2020 Lunar New Year in China [1] and rapidly led to a global spread. The World Health Organization (WHO, Geneva, Switzerland) declared that the new coronavirus outbreak is an international public health concern on 30 January 2020 [2] and WHO officially announced COVID-19 as a pandemic on 11 March 2020 [3]. On 16 March 2020, the US federal government announced the '30 Day to Slow' guideline in response to the pandemic. While implementation of mitigation measures varied by US state and district, many of those declarations were announced in March [4]. For example, Maryland enacted a "Prohibiting Large Gatherings and Events and Closing Senior Centers" order on 12 March 2020; a statewide "stay-at-home" order went into effect on 30 March [5]. California declared a state of emergency in March and a stay-at-home order in 19 March. Under these orders, residents were permitted to go outside "for fresh air and exercise as long as they are maintaining a safe distance from others."

Social distancing, or physical distancing, has widespread consequences, affecting the economy and individuals' behaviors in various ways [6], including significant decreases in human mobility and

traffic volume around the period when various governments announced interventions (i.e., guidelines for social distancing, quarantine, and stay-at-home orders) [1,7,8]. While the definition of mobility varies among studies and disciplines, mobility analysis examines possible travel destinations and travel route based on local land use and demographics [9]. A better understanding of the mobility (e.g., destinations) of people can assist decision-making in prevention of disease transmission [10]. For this, studies showed that aggregated human mobility based on mobile phone data can assist studies assessing the spread of epidemics [11], economic consequence of the COVID-19 pandemic [6,12], and how the stay-at-home orders are effective to mitigate human mobility and thereby reduce the COVID-19 transmission [1,13,14]. Large mobility reduction was detected following the COVID-19 pandemic and specific government directives in the US and globally [15–17]. A study using mobility data from Wuhan and transmission of cases across China found that the positive relationship between human mobility and COVID-19 cases decreased after control measures [1]. A few other studies also suggested that sustained human mobility due to domestic and/or international air travel bans contributed to decreased transmission of COVID-19 cases at the early stages of the outbreak in European countries [18] and China [19]. Given the clear links between mobility and spread of the novel coronavirus [20], understanding mobility patterns is crucial to address COVID-19 outbreaks and develop policies to minimize transmission. While human mobility data have been utilized in some scientific works regarding visualizing mobility patterns [21] and economic effects of mobility changes [6], little is known about whether and how environmental factors influence human mobility under normal conditions and during the COVID-19 pandemic.

Alongside government directives for staying at home and social/physical distancing, health authorities including U.S. Centers for Disease Control and Prevention (CDC) andWHO emphasized the importance of regularly performing exercise to cope with the stress of quarantine, stay healthy, and maintain immunity [22]. Several studies argued that physical inability as a consequence of strict quarantine may be associated with increased risk of mental health outcomes as well as cardiovascular diseases, metabolic diseases, and cancer [23–26]. Thus, the COVID-19 pandemic along with other major environmental crises such as climate change shed a light on the need for better understanding of how to promote resilience or capacity of societies to deal with complex health crises [26].

Earlier work indicates that green space provides health benefits [25,27] and sustainability in cities [28]. Green space is defined as natural vegetation such as grass, bush, plants or trees and the built green structures such as parks and unstructured vegetated areas [29]. Potential pathways for the health benefits from green space include encouraging physical activities and providing direct interactions with nature [30]. Given the restrictions on the gathering of people particularly in indoor settings during COVID-19, understanding the use of green space contributes to our understanding of how green space relates to the ability of communities to cope with the stress from quarantine and pandemic, such as by playing a role as an alternative place for physical activity. A recent study in Oslo, Norway found that outdoor physical activity levels increased after the lockdown was implemented, and that the increases were highest in trails with greener and more remote areas [31]. A study conducted in the US found that the reduction in mobility to parks impacted by state-of-emergency declarations was smaller than the mobility reduction for other venues across the states [32]. Thus, green space could be an effective modifier on the effectiveness of COVID-19 mitigation measures, and such measures could indirectly impact the public health benefits of greenness.

As of April 2020, most US states have ordered nonessential businesses such as restaurant, bars, theaters, and gyms to close but the status of green spaces and open spaces such as parks have been far less consistent in many states. Green space may be one of the limited outdoor places where people seek to perform outdoor exercise during the COVID-19 crisis. The aim of this study was to investigate green space as a potential factor influencing mobility patterns during a pandemic. We hypothesized that the expected mobility decreases due to the social/physical distancing and associated policies will be lower in areas with higher local green space. Specifically, for a case study region of Maryland and California, we examined how the temporal trends in the number of people traveling among the study regions (minor civil division (MCDs) for Maryland, census county division (CCD) for California) differ by local vegetation level. The results here provide information relevant for the design and effectiveness of sheltering policies designed to mitigate a pandemic.
