Paul Switzer

A transformation for ordering multispectral data in terms of image quality with implications for noise removal

A transformation known as the maximum noise fraction (MNF) transformation, which always produces new components ordered by image quality, is presented. It can be shown that this transformation is equivalent to principal components transformations when the noise variance is the same in all bands and that it reduces to a multiple linear regression when noise is in one band only. Noise can be effectively removed from multispectral data by transforming to the MNF space, smoothing or rejecting the most noisy components, and then retransforming to the original space. In this way, more intense smoothing can be applied to the MNF components with high noise and low signal content than could be applied to each band of the original data. The MNF transformation requires knowledge of both the signal and noise covariance matrices. Except when the noise is in one band only, the noise covariance matrix needs to be estimated. One procedure for doing this is discussed and examples of cleaned images are presented. >

The National Human Activity Pattern Survey (NHAPS): A resource for assessing exposure to environmental pollutants

Because human activities impact the timing, location, and degree of pollutant exposure, they play a key role in explaining exposure variation. This fact has motivated the collection of activity pattern data for their specific use in exposure assessments. The largest of these recent efforts is the National Human Activity Pattern Survey (NHAPS), a 2-year probability-based telephone survey ( n=9386) of exposure-related human activities in the United States (U.S.) sponsored by the U.S. Environmental Protection Agency (EPA). The primary purpose of NHAPS was to provide comprehensive and current exposure information over broad geographical and temporal scales, particularly for use in probabilistic population exposure models. NHAPS was conducted on a virtually daily basis from late September 1992 through September 1994 by the University of Marylands Survey Research Center using a computer-assisted telephone interview instrument (CATI) to collect 24-h retrospective diaries and answers to a number of personal and exposure-related questions from each respondent. The resulting diary records contain beginning and ending times for each distinct combination of location and activity occurring on the diary day (i.e., each microenvironment). Between 340 and 1713 respondents of all ages were interviewed in each of the 10 EPA regions across the 48 contiguous states. Interviews were completed in 63% of the households contacted. NHAPS respondents reported spending an average of 87% of their time in enclosed buildings and about 6% of their time in enclosed vehicles. These proportions are fairly constant across the various regions of the U.S. and Canada and for the California population between the late 1980s, when the California Air Resources Board (CARB) sponsored a state-wide activity pattern study, and the mid-1990s, when NHAPS was conducted. However, the number of people exposed to environmental tobacco smoke (ETS) in California seems to have decreased over the same time period, where exposure is determined by the reported time spent with a smoker. In both California and the entire nation, the most time spent exposed to ETS was reported to take place in residential locations.

A proposed sampling constant for use in geochemical analysis

The error in a determination of an element in a rock or mineral sample depends on the analytical error, the weight of sample analysed, and the nature and history of the laboratory sample. The most probable result is not independent of the weight of sample analysed. This is due to the fact that trace constituents often reside in isolated mineral grains. The chance of such mineral grains appearing in any one analysed sample becomes more remote as the sample weight decreases, even when rock or mineral samples are reduced to fine powders. Such subsampling errors can be controlled through the use of sampling constants. These may be estimated by several procedures, including repetitive determination of a constituent and physical measurement of relevant sample characteristics. Sampling constants can be usefully employed during the establishment and certification of reference samples or standards. When subsampling is deficient, analytical results may yield erroneously low values, sometimes with high precision. High precision never implies high accuracy; it may be a symptom of gross error.

Real–Time Measurement of Outdoor Tobacco Smoke Particles

Abstract The current lack of empirical data on outdoor tobacco smoke (OTS) levels impedes OTS exposure and risk assessments. We sought to measure peak and time-averaged OTS concentrations in common outdoor settings near smokers and to explore the determinants of time-varying OTS levels, including the effects of source proximity and wind. Using five types of real-time airborne particle monitoring devices, we obtained more than 8000 min worth of continuous monitoring data, during which there were measurable OTS levels. Measurement intervals ranged from 2 sec to 1 min for the different instruments. We monitored OTS levels during 15 on-site visits to 10 outdoor public places where active cigar and cigarette smokers were present, including parks, sidewalk cafés, and restaurant and pub patios. For three of the visits and during 4 additional days of monitoring outdoors and indoors at a private residence, we controlled smoking activity at precise distances from monitored positions. The overall average OTS respirable particle concentration for the surveys of public places during smoking was approximately 30 μg m−3. OTS exhibited sharp spikes in particle mass concentration during smoking that sometimes exceeded 1000 μg m−3 at distances within 0.5 m of the source. Some average concentrations over the duration of a cigarette and within 0.5 m exceeded 200 μg m−3, with some average downwind levels exceeding 500 μg m−3. OTS levels in a constant upwind direction from an active cigarette source were nearly zero. OTS levels also approached zero at distances greater than approximately 2 m from a single cigarette. During periods of active smoking, peak and average OTS levels near smokers rivaled indoor tobacco smoke concentrations. However, OTS levels dropped almost instantly after smoking activity ceased. Based on our results, it is possible for OTS to present a nuisance or hazard under certain conditions of wind and smoker proximity.

Air change rates of motor vehicles and in-vehicle pollutant concentrations from secondhand smoke

The air change rates of motor vehicles are relevant to the sheltering effect from air pollutants entering from outside a vehicle and also to the interior concentrations from any sources inside its passenger compartment. We made more than 100 air change rate measurements on four motor vehicles under moving and stationary conditions; we also measured the carbon monoxide (CO) and fine particle (PM2.5) decay rates from 14 cigarettes smoked inside the vehicle. With the vehicle stationary and the fan off, the ventilation rate in air changes per hour (ACH) was less than 1 h−1 with the windows closed and increased to 6.5 h−1 with one window fully opened. The vehicle speed, window position, ventilation system, and air conditioner setting was found to affect the ACH. For closed windows and passive ventilation (fan off and no recirculation), the ACH was linearly related to the vehicle speed over the range from 15 to 72 mph (25 to 116 km h−1). With a vehicle moving, windows closed, and the ventilation system off (or the air conditioner set to AC Max), the ACH was less than 6.6 h−1 for speeds ranging from 20 to 72 mph (32 to 116 km h−1). Opening a single window by 3″ (7.6 cm) increased the ACH by 8–16 times. For the 14 cigarettes smoked in vehicles, the deposition rate k and the air change rate a were correlated, following the equation k=1.3a (R2=82%; n=14). With recirculation on (or AC Max) and closed windows, the interior PM2.5 concentration exceeded 2000 μg m−3 momentarily for all cigarettes tested, regardless of speed. The concentration time series measured inside the vehicle followed the mathematical solutions of the indoor mass balance model, and the 24-h average personal exposure to PM2.5 could exceed 35 μg m−3 for just two cigarettes smoked inside the vehicle.

Extensions of linear discriminant analysis for statistical classification of remotely sensed satellite imagery

Linear discriminant analysis is a commonly used statistical tool for the classification of surface features using satellite surface reflectance data. Extensions of this basic tool promise substantial improvements. In particular, we examine the added effectiveness of the integration of spatial autocorrelation into the discriminant model, the resolution of nonhomogeneous pixels, and data based prior probability estimates of class membership.

Graphical Assessment of Dependence: Is a Picture Worth 100 Tests?

Nonparametric tests of association have been around for a long time, and new ones continue to be proposed to cope with specie c forms of association. However, graphs have the potential to assess a far richer class of bivariate dependence structures than any collection oftests. This article describes howchi-plots, used in conjunction with the usual scatterplot, provides a useful practical tool in this regard.Nonparametric tests of association have been around for a long time, and new ones continue to be proposed to cope with specific forms of association. However, graphs have the potential to assess a far richer class of bivariate dependence structures than any collection of tests. This article describes how chi-plots, used in conjunction with the usual scatterplot, provides a useful practical tool in this regard.

Investigations of the proximity effect for pollutants in the indoor environment.

More than a dozen indoor air quality studies have reported a large discrepancy between concentrations measured by stationary indoor monitors (SIMs) and personal exposure monitors (PEMs). One possible cause of this discrepancy is a source proximity effect, in which pollutant sources close to the respondent cause elevated and highly variable exposures. This paper describes three sets of experiments in a home using real-time measurements to characterize and quantify the proximity effect relative to a fixed distant location analogous to a SIM. In the first set of experiments, using sulfur hexafluoride (SF6) as a continuously emitting tracer pollutant from a point source, measurements of pollutant concentrations were made at different distances from the source under different air exchange rates and source strengths. A second set of experiments used a continuous point source of carbon monoxide (CO) tracer pollutant and an array of high time resolution monitors to collect simultaneous concentration readings at different locations in the room. A third set of experiments measured particle count density and particle-bound polycyclic aromatic hydrocarbon (PAH) concentrations emitted from a continuous particle point source (an incense stick) using two particle counters and two PAH monitors, and included human activity periods both before and during the source emission period. Results from the SF6 and CO experiments show that while the source is emitting, a source proximity effect can be seen in the increases in the mean and median and in the variability of concentrations closest to the source, even at a distance of 2.0 m from the source under certain settings of air exchange rate and source strength. CO concentrations at locations near the source were found to be higher and more variable than the predictions of the mass balance model. For particles emitted from the incense source, a source proximity effect was evident for the fine particle sizes (0.3 to 2.5 µm) and particle-bound PAH up to at least 1.0 m from the source. Analysis of spatial and temporal patterns in the data for the three tracer pollutants reveal marked transient elevations of concentrations as seen by the monitor, referred to as “microplumes,” particularly at locations close to the source. Mixing patterns in the room show complex patterns and directional effects, as evidenced by the variable intensity of the microplume activity at different locations. By characterizing the spatial and temporal variability of pollutant concentrations in the home, the proximity effect can be quantified, leading to improved indoor monitoring designs and models of human exposure to air pollutants.

Reassessing the relationship between ozone and short-term mortality in U.S. urban communities

Time-series studies that use daily mortality and ambient ozone concentrations exhibit estimates of ozone effects that are variable across cities. We investigate this intercity variability, as well as the sensitivity of the ozone- mortality associations to modeling assumptions and choice of daily ozone metric, based on reanalysis of data from the National Morbidity, Mortality and Air Pollution Study (NMMAPS). Previous work from NMMAPS reported a statistically significant association between ambient 24-h ozone and short-term mortality when averaged across 98 U.S. cities. Separation of ozone health associations from effects due to weather and co-pollutants is central to their interpretation. We examined the sensitivity of city-specific ozone-mortality estimates to adjustments for confounders and effect modifiers, showing substantial sensitivity. We examined ozone-mortality associations in different concentration ranges, finding a larger incremental effect in higher ranges, but also larger uncertainty. Alternative ozone exposure metrics defined by maximum 8-h averages or 1-h maxima show different ozone-mortality associations that cannot be explained by simple scaling relationships. The emphasis in earlier studies based on NMMAPS has been on the reporting of “national” effects, together with prediction intervals that suggest that these national values are precisely estimated. Our view is that ozone-mortality associations, based on time-series epidemiologic analyses of daily data from multiple cities, reveal still-unexplained inconsistencies and show sensitivity to modeling choices and data selection that contribute to serious uncertainties when epidemiological results are used to discern the nature and magnitude of possible ozone-mortality relationships or are applied to risk assessment.

Observed 1970-2005 cooling of summer daytime temperatures in coastal California.

Abstract This study evaluated 1950–2005 summer [June–August (JJA)] mean monthly air temperatures for two California air basins: the South Coast Air Basin (SoCAB) and the San Francisco Bay Area (SFBA). The study focuses on the more rapid post-1970 warming period, and its daily minima temperature Tmin and maxima temperature Tmax values were used to produce average monthly values and spatial distributions of trends for each air basin. Additional analyses included concurrent SSTs, 40-yr European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analysis (ERA-40) sea level coastal pressure gradients, and GCM-downscaled average temperature Tave values. Results for all 253 California National Weather Service (NWS) Cooperative Observer Program (COOP) sites together showed increased Tave values (0.23°C decade−1); asymmetric warming, as Tmin values increase faster than Tmax values (0.27° versus 0.04°C decade−1) and thus decreased daily temperature range (DTR) values (0.15°C decade−1). The spatial distribution of...