Background
Parks are venues for physical activity and most localities maintain a park system to serve their constituents. Park use by different gender, age, and ethnic/racial groups has not been well-studied and how much parks contribute to population-level physical activity is unknown.

Objectives
1) To characterize park use, including user characteristics and physical activity mode and intensity.
2) To determine how local residents interact with parks based upon their proximity, size and features.
3) To estimate the population-level impact of local parks on physical activity.

Methods
We studied the use of 16 local parks in Los Angeles, including 12 neighborhood parks, two skate parks and two senior citizens centers located within larger parks. We used the System for Observing Play and Recreation in Communities (SOPARC) in each park during four one-hour time periods per day (7:30 a.m.; 12:30 p.m., 3:30 p.m., 6:30 p.m.) during seven days of clement weather. We coded the mode and activity level (i.e., sedentary, walking, vigorous) of approximately 30,000 park users and characterized them by gender, age group (child, teen, adult, senior), and race/ethnicity (Latino, African-American, white and other). We also conducted intercept interviews with 1,316 adult park users and interviewed 1,214 adults in randomly selected homes within a quarter-, half-, one-, or two- mile buffers around each park.

We geocoded the residential address of both park users and residents and calculated their proximity to the park. We also estimated the number of people living within a one-mile radius of each park from the U.S. 2000 Census data. Using two methods, direct observations of park users and self-reports of parks use by the random sample of residents, we estimated the total number of people who used the parks. Using regression analysis, we modeled park use and frequency of exercise controlling for individual (i.e., age, gender, race/ ethnicity) and park characteristics.

Results
On average, the 12 neighborhood parks served approximately 67,300 persons in a one-mile radius with an average of 0.16 acres/1,000 population. We observed an average of 2,180 persons per park during the week of observation (range 646-4,857). The age distribution was 30% children, 15% teens, 47% adults and 7% older adults (over 60). The ethnic/racial distribution was 67% Latino, 12% white, 14% African American, and 8% Asian or other. Among observed park users, 61% were sedentary (range = 46%-76%), 23% walking (range = 12%-35%), and 16% vigorous (range = 11%-23%). Males were nearly twice as likely to be engaged in vigorous activity as females (19% versus 12%) and were seen in parks in greater numbers than females (63% versus 37% of users). In one week, we observed an average of 356 seniors in each senior center; an average of 102 children and 74 teens in each of two skate parks (84% male); an average of 156 children and 16 teens in each playground; and an average of 305 adults using each track. Organized activities attracted an average of 41 persons per event compared to three persons for unorganized activities in the same location. Survey results indicate the best predictor of park use was residential proximity to the park. Proximity also predicted frequency of exercise: 42% residents living within a half-mile of the park reported exercising >5x/week, compared to 25% of those who lived greater than one-mile away (p<.002). The local park was identified by 52% of residents as the most common place for exercise and by all ethnic/racial groups. Twenty-five percent of whites reported using health clubs, compared to 21% of Asian and 8% of Latino and African American respondents. Organized activities appear to account for large variations in the number of park users as well as levels of energy expenditure. METs (number of users times the observed physical activity levels) expended varied considerably by park (mean = 5,169; range = 1,524-10,094), by park user (mean = 2.6, range = 2.2-2.9), by gender (males = 2.7 and females = 2.3) and by population in a one-mile radius (mean = .19, range = .08-.34). Estimates of park use each week by self-report and observation differed significantly, with self-reported use being an average of 4.7 times higher that what we observed. On average, activity areas in the park were empty 53% of the time.

Conclusions

Parks serving similar populations differ dramatically in their use and how active people are in them. People appear to overestimate their use of the park and parks are underutilized. Proximity is the main determinant of individual use. Because park characteristics differ considerably, studies of a larger number of parks are needed to more fully understand their relationship with population use and physical activity.

Support
National Institute of Environmental Health Sciences #P50ES012383.