Lidar Applications in Air Pollution Monitoring
 

Investigators

William Eichinger (University of Iowa ),   Daniel Cooper (Los Alamos National Laboratory), Larry Tellier (Los Alamos National Laboratory), Howard Kraye (Santa Fe Technology)

Sponsor

New Jersey Department of Transportation

Measurement of Traffic Particulate Emissions and Incident Detection

Improving the environment is an important objective in the design and development of Intelligent Transportation Systems (ITS). The Clean Air Act Amendment of 1991 (CAA) and the Intermodal Surface Transportation Efficiency Act (ISTEA) have placed more demands on metropolitan areas to develop surface transportation plans that will improve air quality. The State of New Jersey is installing an Incident Management System (IMS) in order to mitigate the effects of traffic accidents and incidents on traffic flow. The objective of this project is to determine the environmental impact of the application of technologies for congestion management to urban transportation systems by making measurements of air quality before and after the installation of the system. This is done using lidar (LIght Detection And Ranging) technology in with conventional point instruments.

Experiment Site and Equipment Locations

Results

The second graph above shows the results of a new technique to measure the emission rate of Carbon Monoxide from vehicles on the highway. It uses an atmospheric similarity model using temperatures and CO concentrations made at two altitudes.

Three case studies were done using the lidar data to demonstrate that traffic incidents can be clearly identified using lidar backscatter returns and comparing the abnormally high concentrations of particulates as compared to the concentrations found in the air nearby. If one plots the particulate concentration (measured as the corrected lidar signal) over the highway versus the concentration of the nearby air mass but not over a particulate source, one obtains the figure to the left. This figure is a composite made from all three of the case studies showing the difference between roadway concentrations during normal and incident levels of traffic as compared to the ambient air concentration. One can see that under normal traffic, the particulate concentration over the roadway is about 35% higher than the ambient air while the concentration during an incident is three to ten times the concentration of the ambient air and is above a threshold value of about 10.

This allows a simple method of discriminating between normal and incident traffic on the basis of the particulate concentration over the roadway which can be measured remotely. This simple method has been shown to work in poor weather situations in which the weather was intermittently foggy with low level clouds. In this case, the ambient air approaches those levels of concentration (caused mostly by the foggy conditions) normally found during incidents. These criterion may serve as a clear indicator for incidents; one that could easily be programmed into an automated sensor network. Another conclusion that can be drawn is that incidents are a major contributor to local air pollution. Since the concentration above a source is proportional to the source strength, it follows that the source is three to ten times as great during an incident as compared to normal traffic. It is likely that this reasoning holds for all of the primary pollutants, CO, NO, and NO2.

The data shown here imply that the emission rate of particulates is several times higher during incidents. This conclusion is also borne out by studies showing that the amount carbon monoxide emitted per mile decreases dramatically with increasing speed [Sturm et al, 1997]. NOx emissions also tend to decrease, but with far more spread in the data. Also significant is that a major factor in the emission rates for automobiles is the presence or absence of acceleration. Changing speed (a condition often found in incidents) generally results in increasing amounts of pollutants.

If the emission rates of the various pollutants are roughly proportional, and since incidents have 3 to 10 times the emission rates of normal traffic, it stands to reason that reducing the length and impact of incidents, will help reduce local pollution concentrations. Further, any measures that are taken to keep traffic flowing smoothly will help reduce the pollutant concentrations. The amount of decrease in pollution will be dependent upon the fraction of time a particular stretch of highway experiences an incident. In some areas that are lightly trafficked, the benefit may not be great. However, those areas with high incident rates are most often those that are highly traveled. Thus reductions in pollution will occur in those areas where they are needed most.

Measurement of Traffic Particulate Emissions and Incident Detection

The plots of lidar data that are displayed are graphical depictions of the intensity of the lidar return signal in two dimensions. The signals have been corrected for range and average attenuation effects. It has been been shown that the backscatter signal from a 1.064 micron lidar tracks the concentration of particulates with diameters in the range of 0.5 to 2.5 microns. Thus while the data shown in the three case studies is not a direct measure of the concentration of particulates, the variations in the lidar return do represent variations in the overall particulate concentrations. Recognizing that there are several limitations to this approach, the lidar data are treated as proportional to the particulate concentrations and the plots as maps of the relative particulate concentrations.

The intensities of the lidar returns are depicted as colors in the plot. Higher intensities that correspond to higher particulate concentrations are shown as red. Similarly, lower intensities that correspond to lower particulate concentrations are shown in blue.

One of the advantages of using the lidar is that it can measure out to long distances (10 km) with high spatial resolution. A disadvantage is that images on that scale often do not show small scale features clearly. The image above is an example of a 5 km lidar scan shown with the same vertical scale as horizontal. The scan to the right is also a vertical slice of the atmosphere, but the vertical scale is exaggerated to show details near the surface. While this may lead to distorted structures (for example the plume at 3000 m), it enables the identification of numerous structures near the surface that can be associated with roads and intersections.

High Resolution Capabilities

Above: An example of a horizontal scan showing the effluents from cars backed up along US 46 in Northern New Jersey. The bright line starts at a major on ramp (at about 2000 m) and extends down the highway (to the right).

Left: A blowup of a portion of a vertical lidar scan showing the particulate plumes from the two portions of the highighway. This scan has a 7.5 meter range resolution. Currently the system is capable of 1.5 meter resolution so that detailed maps of the fine scale resolution are possible. Maps such as these, when combined with some local meteorological information can provide estimates of the emission rates of particulates from the highway.

Results

The graphs above show examples of incidents in which traffic was stopped or slowed, leading to abnormally high concentrations of particulates in the air immediately above the highway. The graph on the upper right is a typical condition showing normal traffic volume and speed. This should be contrasted to the other three.

It is clear that the presence or absence of incidents is clearly seen by the lidar. From the data accumulated by this project, a simple algorithm can be used to detect the location and the occurrence of incidents that delay traffic. Another conclusion that can be drawn is that these incidents are a major contributor to local air pollution. Since the concentration above a source is proportional to the source strength, it follows that the source is three to ten times as great during an incident as compared to normal traffic. It is likely that this reasoning holds for all of the primary pollutants, such as CO, NO, and NO2 as well as for the particulates.

An ongoing area of research is the effort to invert the measured concentration values to obtain surface parameters and the source strength (the emission rate). Mathematical models are available which predict the shape of these plumes in closed form. Unfortunately, these functions are extremely difficult to work with. Should the effort succeed, a simple measurement of wind speed at one altitude combined with the spatial distribution of the pollutant will provide valuable information for modelers and planners.