Background, aim, and scope: Ten years of public health interventions on industrial emissions to clean air were monitored for the Mediterranean city of Cartagena. During the 1960s, a number of large chemical and non-ferrous metallurgical factories were established that significantly deteriorated the city's air quality. By the 1970s, the average annual air concentration of sulfur dioxide (SO2) ranged from 200 to 300 microg/m3 (standard conditions units). In 1979, the Spanish government implemented an industrial intervention plan to improve the performance of factories and industrial air pollution surveillance. Unplanned urban development led to residential housing being located adjacent to three major factories. Factory A produced lead, factory B processed zinc from ore concentrates, and factory C produced sulfuric acid and phosphates. This, in combination with the particular abrupt topography and frequent atmospheric thermal inversions, resulted in the worsening of air quality and heightening concern for public health. In 1990, the City Council authorized the immediate intervention at these factories to reduce or shut down production if ambient levels of SO2 or total suspended particles (TSP) exceeded a time-emission threshold in pre-established meteorological contexts. The aim of this research was to assess the appropriateness and effectiveness of the intervention plan implemented from 1992 to 2001 to abate industrial air pollution.

Materials and methods: The maximum daily 1-h ambient air level of SO2, NO2, and TSP pollutants was selected from one of the three urban automatic stations, designed to monitor ambient air quality around industrial emissions sources. The day on which an intervention took place to reduce and/or interrupt industrial production by factory and pollutant was defined as a control day, and the day after an intervention as a post-control day. To assess the short-term intervention effect on air quality, an ecological time series design was applied, using regression analysis in generalized additive models, focusing on day-to-day variations of ambient air pollutants levels. Two indicators were estimated: (a) appropriateness, the ratio between mean levels of the pollutant for control days versus the other days, and (b) effectiveness, the ratio between mean levels of the pollutant for post-control days versus the other days. Ratios in regression analyses were adjusted for trend, seasonality, temperature, humidity and atmospheric pressure, calendar day, and special events as well as the other pollutants.

Results: A total of 702 control days were made on the factories' industrial production during the 10-year period. Fifteen reductions and five shutdown control days took place at factory A for ambient air SO2. At factory B, more controls were carried out for the SO2 pollutant in the years 1992-1993 and 1997. At factory C, the control days for SO2 decreased from 59 reductions and 14 shutdowns to a minimum from 1995 onwards, whereas the controls on TSP were more frequent, reaching a maximum of 99 reductions and 47 shutdowns in the last year. SO2 ambient air mean levels ranged from 456 to 699 microg/m(3) among factories on reduction control days and between 624 and 1,010 microg/m(3) on shutdown days. The TSP ambient air mean levels were 428 and 506 microg/m(3) on reduction and shutdown days, respectively. For all types of control days and factories, a mean ratio of 104% (95% confidence interval [CI] 88 to 121) in SO(2) levels was obtained and a mean ratio of 67% (95% CI 59 to 75) in TSP levels. Post-control days at all factories showed a mean ratio of -16% (95% CI -7 to -24) in SO(2) levels and a mean ratio of -13% (95% CI -7 to -19) in TSP levels.

Discussion: Interventions on industrial production based on the urban SO(2) and TSP ambient air levels were justified by the high concentrations detected. The best assessment of the interventions' effectiveness would have been to utilize the ambient air pollutant concentration readings from the entire time of the production shutdowns or reductions; however, the daily hourly maximum turned out to be a useful indicator because of meteorological factors influencing the diurnal concentration profile. A substantial number of interventions were carried out from 1 to 3 AM: , when vehicular traffic was minimum. On the other hand, atmospheric stability undergoes diurnal cycling in the autumn-winter period due to thermal inversion, which reaches maximum levels around daybreak. Therefore, this increases the ambient air levels and justified the interventions carried out at daybreak in spite of the traffic influence.

Conclusions: All the interventions for SO(2) and TSP were carried out when the measured ambient air levels of pollutants were exceeded, which shows the appropriateness of the intervention program. This excess was greater when intervening on SO(2) than on the TSP levels. For both ambient air levels of SO(2) and TSP, significant drops in air pollution were achieved from all three factories following activity reductions. The production shutdown controls were very effective, because they returned excess levels, higher than in the reduction controls, to everyday mean values.

Recommendations and perspectives: The Cartagena City observational system of intermittent control has proven to effectively reduce industrial emissions' impact on ambient air quality. This experienced model approach could serve well in highly polluted industrial settings. From a public health perspective, studies are needed to assess that the industrial interventions to control air pollution were related to healthier human populations. Legislation was needed to allow the public administration to take direct actions upon the polluting industries.