Creel surveys are valuable tools in recreational fisheries management. However, multiple‐impoundment fisheries of complex spatial structure can complicate survey designs and pose logistical challenges for management agencies. Marben Public Fishing Area in Mansfield, GA is a multi‐impoundment fishery with many access points, and these features prevent or complicate use of traditional on‐site contact methods such as standard roving‐ or access‐point designs because many anglers may be missed during the survey process. Therefore, adaptation of a traditional survey method is often required for sampling this special case of multi‐lake fisheries to develop an accurate fishery profile. Accordingly, a modified non‐uniform probability roving creel survey was conducted at the Marben PFA during 2013 to estimate fishery characteristics relating to fishing effort, catch, and fish harvest. Monthly fishing effort averaged 7,523 angler‐hours (h) (SD = 5,956) and ranged from 1,301 h (SD = 562) in December to 21,856 h (SD = 5909) in May. A generalized linear mixed model was used to determine that angler catch and harvest rates were significantly higher in the spring and summer (all p < 0.05) than in the other seasons, but did not vary by fishing location. Our results demonstrate the utility of modifying existing creel methodology for monitoring small, spatially complex, intensely managed impoundments that support quality recreational fisheries and provide a template for the assessment and management of similar regional fisheries.