This study describes the characteristics and physical activity levels of park visitors in two large metropolitan parks in Melbourne, Australia. Given that physical inactivity is a major contributor to the burden of chronic disease, including cardiovascular disease, diabetes, and overweight and obesity [27], understanding the characteristics of park visitors and park visitation is important in order to develop strategies to increase park use and park-based physical activity.

Across all observations, 62 % of park visitors were observed lying, sitting or standing. This is an important observation and shows that visiting large metropolitan parks is not always for physical activity purposes. This is consistent with observational studies of park visitation in the US reporting that more than half of park visitors engage in sedentary behavior (primarily sitting) during their park visit [9, 12]. A previous Australian study found that compared with poorly designed parks, more people used attractive parks, and that visitors engaged in a wider range of activities, including sedentary behaviors such as picnics and sitting [5]. The parks included in the current study are large attractive metropolitan parks with amenities that attract visitors from outside the neighborhood and which encourage picnics and social gatherings. These activities are more likely to involve sedentary behaviors. Previous research has shown that visitors who travel greater distances to a park tend to visit for picnics and sedentary social activities with family [28]. Although spending time in parks may benefit mental health [29], as such a large percentage of park visitors were engaging in low levels of physical activity, it is important to better understand how to encourage park users to engage in physical activity while they are in the park. Simply offering attractive large parks may be insufficient to increase already low community-wide physical activity levels.

Some between-park differences in park visitation were observed. More children and seniors were observed in the park located in the low SES area compared with the park in the high SES area and almost seven in ten park visitors were observed lying, sitting or standing in the park in the low SES area, compared with only 57 % of visitors engaging in these behaviors in the park in the high SES area. In addition, park visitors at the park in the high SES area had greater odds of engaging in MVPA and a higher proportion were observed engaged in vigorous activity compared with park visitors in the low SES area. This is in contrast to previous research that showed visitors of low-income neighborhood parks to be more vigorously active than visitors of high-income neighborhood parks [30], although the overall park activity-related intensity in that study did not differ between high- and low-income neighborhoods.

It is possible that the differences in park visitation observed in the current study between the two parks may be reflected by different demographic profiles of residents living near the two parks. Considering that residents living in low SES areas are more likely to be at an increased risk of inactivity and associated poor health outcomes [31], it is important to understand why there may be differences in park-based physical activity between visitors to parks in a low and a high SES area. Previous research in Melbourne, Australia found that parks in low SES areas had fewer amenities and features likely to promote physical activity among children than parks in higher SES areas [32]. Although it is acknowledged that the data used in the current study were collected prior to the development of a new playground at the park in the low SES area, the comparison park in the high SES area was selected based on having similar features and amenities to that of the intervention park at baseline. It is therefore possible that the lower level of MVPA observed in the park in the low SES area may be due to cultural or social factors as the demographic profile of residents in this area is diverse with a growing indigenous population and a high proportion of residents born overseas (49.6 %) [25]. Longitudinal research has shown that having access to green space near home may not be adequate for initiating regular walking habits; however, it seems to be important for maintenance of walking habits [33]. Additional factors such as within park programming may be required to stimulate adults to take up walking or other forms of physical activity in parks. This may be particularly important in low SES areas where visitors may be less likely to be physically active when visiting the park.

The majority of park visitors in the current study were observed on weekend days, but there were no differences in the odds of being observed engaging in MVPA between weekdays and weekend days. Time of day did appear to be important; however, as park visitors were more physically active in the mornings. Previous studies have also found the morning to be a key period for park-based physical activity, with fewer visitors observed being sedentary and more visitors observed walking [30] and engaged in vigorous activity [20] in the morning compared with other times of the day. Thus, it may be important for park programmers and future intervention studies to consider the time of the day and maximise opportunities in the morning when park visitors are most likely to be active, as well as developing strategies to increase activity in the other less active periods of the day. Females were less likely than males to be observed in MVPA, so future research may also consider exploring opportunities for women to engage in park-based physical activity.

It is important to acknowledge the study limitations. Firstly, studies involving direct observations provide contextually-rich information about settings in which physical activity occurs; however, they are a snapshot in time that provide a general indication of park visitation on specified days and demographic characteristics, such as age, are estimated based on best guess. Additional methodological limitations include the cross-sectional design, inclusion of only two parks, and the completion of observations in the one season which limits the ability to generalize the results to other parks and across seasons. It is also important to acknowledge some inconsistencies in the number of observation scans included in the coding of periods of the day. For example, the midday period included two scan periods (12.30 pm and 1.30 pm) whereas the afternoon period included three scan periods (2.30 pm, 3.30 pm and 4.30 pm). Despite these limitations, the study involved a large number of observations which were conducted from early morning until late afternoon on both weekdays and weekend days, the reliability data collected showed high inter-observer agreement which provides adequate confidence in the measures used, and there was consistency in the measures with observations completed at the same time and day at both parks.