A Regression Model to Estimate Baseline Use of a Recreation Area Following an Oil Spill

ABSTRACT Recreation losses resulting from an oil spill are based on the difference between the recreation use that would have occurred in the absence of the spill (i.e., baseline use) and the recreation use that actually occurs following the spill, (i.e., with-injury use). Thus, recreation losses end when actual recreation use returns to baseline levels. Since baseline recreation use is unobserved, by definition, estimating baseline recreation use is an important, and usually controversial, aspect of assessing recreation losses. This paper presents the authors' approach for estimating baseline recreation use of a coastal state park in Texas following a February 1995 oil spill in the Gulf of Mexico. They had daily visitation estimates for the park for all of February and March in 1994 and 1995. Unfortunately, the weather during these months in 1995 was colder, wetter, and windier than the weather in 1994. Thus, the authors could not use the 1994 visitation estimates as a proxy for 1995 baseline visitation at the state park because the 1995 baseline visitation would have been lower than 1994 visitation as a result of poorer weather conditions. Since the authors could not use 1994 visitation estimates as a proxy for 1995 baseline visitation, they developed a regression model to estimate 1995 baseline visitation as a function of variables such as weather conditions (e.g., temperature, amount of sunshine, and precipitation), day of the week, week in the period, and oil-spill cleanup days. The results of their regression model indicate that visits to the state park returned to baseline levels about a week after the cleanup was completed. Their model reveals that almost half of the reduction in 1995 visits to the state park was a result of poor weather conditions, not the oil spill. Finally, the model also suggests that some of the forgone visits immediately following the oil spill were simply postponed, rather than lost.