Abstract

Handheld sun photometers are typically used to make aerosol optical depth measurements while on the ground. Various investigators, in unrelated efforts, have used handheld sun photometers to make aerosol optical depth measurements from light aircraft, but the strengths and weakness of this approach have not been characterized until now. While the ease and relatively low cost of an aircraft manual sun photometer are attractive, determining if the sun photometer was correctly pointed at the sun for each measurement is the biggest challenge. This problem can be partially addressed by collecting a large number of measurements at each altitude, then manually removing the largest optical depths (misalignment always results in erroneous larger values). Examples of past aircraft manual sun photometer measurements are demonstrating that it is possible to obtain quantitative measurements if sufficient sun photometer measurements are made at each elevation. In order to improve on manual sun photometer measurements, a small webcam was attached to the side of a Microtops sun photometer, and the Microtops sun photometer was triggered by computer control. By detecting the position of the sun in the webcam image, it is possible to determine whether the sun photometer was pointed at the sun correctly when the aerosol optical depth measurement was made. Unfortunately, it was found that the Microtops sun photometer takes ∼1.1 s to scan over the five wavelength channels. This 1.1-s delay proved to be too long, preventing the proposed approach from working as the aircraft was bouncing around.

1. Introduction

Aerosol optical depth is one of the key parameters required for radiative transfer calculations and provides a simple way of measuring aerosol loadings in the atmospheric column. Consequently, it is frequently reported and employed in climate studies (Penner et al. 2002), satellite validation (Ignatov and Stowe 2002), emission calculations (Porter et al. 2002), and regional aerosol loading studies (Chin et al. 2002).

Aerosol optical depth measurements have been made for many years with two principal techniques. One approach uses a narrow field of view radiometer (Voltz type) pointed directly at the sun (Voltz 1959; Shaw 1983). For years handheld Voltz-type sun photometers (VSPs), using narrow field of view sensors, have been popular. They are relatively inexpensive and can be pointed at the sun manually (Shaw 1983; Morys et al. 1996, among others). VSPs typically have a field of view of 1°–3° and can provide accurate measurements of the aerosol optical depths. Most VSPs are handheld and produce reliable measurements when measurements are taken from a stable platform, for example, land. Due to need for measurements in ocean regions, VSPs are also routinely used on ships, although with added uncertainty due to platform movement, which causes aiming errors (Porter et al. 2001; Frouin et al. 2001; Knobelspiesse et al. 2003). Automated VSPs, which track the sun using two axis scanners, are also common over land (e.g., Holben et al. 1998). This approach has also been applied to ship (Korotaev et al. 1993) and aircraft measurements (see http://geo.arc.nasa.gov/sgg/AATS-website/; Matsumoto et al. 1987; Redemann et al. 2006).

A second approach to measure aerosol optical depths uses a shadowband radiometer to measure the total and diffuse radiation from which the direct radiation is derived (Harrison et al. 1994). Compared to VSPs this approach is considered slightly less accurate due to blocking of part of the sky and the imperfect cosine response of the detector. This approach does not suffer from pointing challenges, but it does in theory require the system to be level. In the case of Reynolds et al. (2001), the system cosine response was known to a sufficient accuracy that it was possible to correct for the ship pitch, roll, and heading. The shadowband approach also has the advantage of providing simultaneous measurement of downwelling diffuse irradiance. Shadowband-based sun photometers have been implemented on land (Harrison et al. 1994), ship (Miller et al. 2004), and aircraft (Valero et al. 1989).

Small aircraft can be excellent platforms to carry out atmospheric research because of their lower operating costs, their ease of use in many parts of the world for short-term hire, and the availability of pilots for light aircraft. Typically costs range from $150 to $850 h−1 depending on the size of the aircraft. The use of handheld VSPs on light aircraft is a logical and economical approach. The methodology and challenges in using manual sun photometers on light aircraft have never, to our knowledge, been reported in scientific literature. Here we describe the use of manual VSPs aboard light aircraft. Examples obtained from several unrelated efforts are given. When possible the aircraft sun photometer measurements are compared with independent measurements.

2. Background of aircraft manual sun photometer measurements

In the past, various independent investigators have used light aircraft to collect manual sun photometer measurements using VSPs. To our knowledge, Shaw (1975) was the first to carry out manual sun photometer measurements from light aircraft. Using a Cessna 180 aircraft he made measurements over Barrow, Alaska, using a custom VSP. It is interesting to review these first aircraft sun photometer measurements as they illustrate some of the initial challenges and how rapidly the field is changing. A battery operated chart recorder along with hand annotations were used to record sun photometer data. Utilization of a chart recorder included the risk of sprayed red ink caused by changing atmospheric pressure. The pistol grip mounted sun photometer was manually pointed out a 2-in.-diameter hole cut in the Plexiglas side window of a Cessna 180 aircraft. Significant effort went into testing interference filters from many different vendors before selecting temperature-cycled interference filters from Barr Associates. These filters were found to have the lowest drift rates and are still used today in many sun photometers. The early sun photometer also employed a very stable amplifier with operational amplifier feedback with 1%–2% accuracy. At relatively low sun angles, ∼24°, typical at Barrow, Alaska, it is possible to get optical depth to an accuracy of ∼0.02 or better. These early aircraft sun photometer measurements showed a distinct aerosol layer (later known as Artic Haze) to be present over the Artic region. Additional research indicated that these pollutants came from anthropogenic sources in eastern Europe and Russia.

More recently, several investigators have employed hand-operated sun photometers to collect aerosol optical depths from aircraft. Beginning in approximately 1995, Antony Clarke and John Porter collected manual aircraft sun photometer measurements near Hawaii to study marine aerosol conditions. These aircraft measurements continued intermittently for the next 10 yr during which general procedures were developed (Porter et al. 2000). In a separate and unrelated effort, Maring et al. (2003) collected manual sun photometer measurements of airborne dust during the Puerto Rico Dust Experiment (PRIDE) campaign near Puerto Rico from a Cessna 172. Most recently, manual aircraft VSP measurements were made in dusty and polluted environments on a light aircraft as part of the United Arab Emirates Unified Aerosol Experiment (UAE2; summer 2004; http://uae2.gsfc.nasa.gov/).

3. Discussion of the use of manual sun photometers on light aircraft

To make manual sun photometer measurements from an aircraft, the aircraft should have an open window with an upward view, that is, a low wing aircraft is desirable. Many aircraft have a small window (approximately 15 cm × 15 cm) next to the pilots’ seats. While we have successfully used this window for manual sun photometer measurements on several occasions (Alaska, PRIDE, and UAE2 measurements), it is certainly difficult to use. Aircraft with a back window (or door) are better suited for manual sun photometer measurements. We have used a Seneca aircraft, a four-seat aircraft with a back door with a window, which typically does not open. To make measurements from this aircraft, we acquired an additional door and removed its window [allowed under Federal Aviation Administration (FAA) regulations]. Prior to each flight the original aircraft door was removed and the modified door was installed (Fig. 1). This procedure took little time (∼10–15 min) and allowed for relatively easy sun photometer measurements from that aircraft. The Chieftain (Fig. 2) is a larger aircraft (eight seater), which is ideal for manual sun photometer measurements. The Chieftain’s rear door is split with half of the door swinging up and the other half swinging down. By removing the upper half of the door (allowed by FAA regulations), manual sun photometer measurements can be easily made (see Figs. 2 and 3). To allow the operator to rest on a flat surface and to mount other sensors, a small saddle platform, with a padded screw clamp, was built for the lower half of the Chieftain window (see Figs. 3 and 4).

Fig. 1.

Picture of the Seneca back door with the window removed. A small platform is also mounted on the window bottom for mounting sensors. The second door has a custom upward- and downward-looking irradiance sensors.

Fig. 1.

Picture of the Seneca back door with the window removed. A small platform is also mounted on the window bottom for mounting sensors. The second door has a custom upward- and downward-looking irradiance sensors.

Fig. 2.

Chieftain aircraft with top half of back door removed.

Fig. 2.

Chieftain aircraft with top half of back door removed.

Fig. 3.

Manual sun photometer measurements being made from the open window of the Chieftain aircraft.

Fig. 3.

Manual sun photometer measurements being made from the open window of the Chieftain aircraft.

Fig. 4.

Saddle, which fit over the top half of the door allowing easy control of the sun photometer.

Fig. 4.

Saddle, which fit over the top half of the door allowing easy control of the sun photometer.

Typically a small boundary layer exists along the side of the plane (2–5 in. wide) where the airflow is slower. Prior to making sun photometer measurements it is useful to manually determine the depth boundary layer where the wind speeds are reduced. Extending the sun photometer only into this low velocity layer makes measurement easier. It is also wise to attach the Microtops to a lanyard to avoid accidentally dropping the instrument. To prevent reaching far out into the air stream, aircraft sun photometer measurements can be most easily made when the sun elevation is not large. Morning, afternoon, or winter season time periods are best. Flight legs should also be arranged, which provide the longest possible time facing the sun.

The largest problem in making sun photometer measurements from a moving platform is determining whether the VSP was correctly pointed at the sun. Depending on the system field of view and design, some amount of pointing error is acceptable without significant error in the sun photometer measurement. But as the pointing error increases, beyond the acceptable value, then the aerosol optical depth measurement will be larger than the correct value (i.e., positive definite error). Porter et al. (2001) and Knobelspiesse et al. (2003) investigated ship-based manual sun photometer measurements and showed that if a large number of aerosol optical depth measurements are rapidly collected then the larger values can be discarded and the smallest values will most likely be correct. In flight it is difficult to consistently point the sun photometer at the sun. This alignment error (positive bias) is evident in sun photometer measurements made in aircraft (see Figs. 5 and 10). When many repetitive measurements are made, reliable measurements cluster at the lowest value. Misaligned sun photometer measurements tail toward larger aerosol optical depths. In cases where many measurements remain after removing the larger values then the remaining measurements are considered more reliable (see Figs. 5 and 10). When relatively few measurements remain after the screening process (e.g., Fig. 5, hours 11.2–11.3), the reliability of the remaining data is less certain. The situation is similar for sun photometer measurements made from ships, which are also moving platforms (Porter et al. 2001; Knobelspiesse et al. 2003). When eliminating larger optical depth measurements, a general rule is to retain measurements that are within ∼20% of the lowest value or within a value of 0.025 when the optical depths are below 0.08 (Porter et al. 2001). Stricter processing is also possible. For well-calibrated instruments, ground-based sun photometer measurements are expected to have an accuracy of ∼0.01 (Porter et al. 2001). For aircraft and ship measurements the expected accuracy depends on the number of data points remaining after filtering data. Although Porter et al. (2001) and Knobelspiesse et al. (2003) placed values on this error (∼0.025), the accuracy is difficult to estimate.

Fig. 5.

Aerosol optical depths collected from the Seneca aircraft near Hawaii, HI. Some of the aerosol optical depth values were off the top of the chart. The aircraft height is shown in solid line. Open circles represent measurements, which were discarded due to misalignment. Solid circles are considered correct values. The range of values observed within ±3 h of the flight data are shown for the Lanai and the Mauna Loa, HI, AERONET sun photometers.

Fig. 5.

Aerosol optical depths collected from the Seneca aircraft near Hawaii, HI. Some of the aerosol optical depth values were off the top of the chart. The aircraft height is shown in solid line. Open circles represent measurements, which were discarded due to misalignment. Solid circles are considered correct values. The range of values observed within ±3 h of the flight data are shown for the Lanai and the Mauna Loa, HI, AERONET sun photometers.

Fig. 10.

Aerosol optical depth measurements obtained during the UAE 2004 experiment. Solid circles are considered good data. The star symbols with the solid line show the average of the good optical depths. When the aircraft altitude did not change the average of values for that altitude are calculated together.

Fig. 10.

Aerosol optical depth measurements obtained during the UAE 2004 experiment. Solid circles are considered good data. The star symbols with the solid line show the average of the good optical depths. When the aircraft altitude did not change the average of values for that altitude are calculated together.

With the exception of the Shaw measurements (Shaw 1975), all of the aerosol optical depth measurements described here were carried out with handheld Microtops sun photometers. The Microtops sun photometer (manufactured by Solar Light, Inc.) measures light in five wavelength bands defined by the interference filters installed. While the Barr filters typically used have a long field life, decay in some filters has been observed (Porter et al. 2001). Routine Langley plot calibration is also required to ensure that the sun photometer is properly calibrated (Shaw 1983; Holben et al. 1998). Depending on the choice of filter wavelengths, the instrument can measure aerosol optical depths, column ozone burdens (Morys et al. 1996), and precipitable water vapor (Ichoku et al. 2002). The Microtops sun photometer has a built in pressure and temperature sensors and allows for a GPS connection to obtain the position and time. The Microtops also carries out a dark calibration upon startup, which corrects for temperature effects. Based on temperature tests, we found that the Microtops should be powered down routinely to update this temperature correction (Porter et al. 2001).

4. Examples of aircraft manual sun photometer measurements

On 19 March 2000, sun photometer measurements were carried out from a Seneca aircraft over the Moby buoy south of Lanai, Hawaii (see Fig. 5). The marine boundary layer was hazy with sea salt due to strong trade wind conditions where wind speeds were often >15 m s−1. The ocean surface was covered with many whitecaps, especially over channels between islands where stronger winds occurred. The aerosol optical depth at 500 nm increased from ∼0.03 above the trade wind inversion to ∼0.17 as the aircraft descended toward the surface (Fig. 5). On this windy day, turbulence was quite high and a large number of the sun photometer measurements were misaligned. Near the surface, only a small set of optical depth measurements had low values, which could be considered correct. The fact that only a few values remained after the screening process means that even these are possibly suspect. The measurements collected from 11.35 to 11.45 are too high in comparison to the lower values at hour 11.5.

At the time of these measurements, Aerosol_Robotic Network (AERONET) sun photometer measurements of the free troposphere at the Mauna Loa Observatory averaged 0.03 and ranged from 0.015 to 0.03 within ±3 h of the aircraft measurements shown in Fig. 5. Measurements from the aircraft were essentially identical at 0.03.

On the lee island of Lanai, AERONET sun photometer measurements of optical depth ranged from 0.11 to 0.14 (at 500 nm) within ±3 h of the aircraft measurements (Fig. 5). Measurements from the aircraft, made at low altitude, averaged significantly higher (0.17) than the AERONET ground-based measurements. This difference may be due to instrument misalignment, as discussed above, but could also be due to stronger winds near the Moby site where the flight descent occurred. Often stronger winds occur in the channel between Maui and Molokai, and the strong wind area can extend very close to the Moby site. These higher wind speeds may have produced locally higher concentrations of sea salt aerosols. During this Seneca flight aerosol size distributions were measured using a forward scattering spectrometer probe (FSSP; Porter et al. 2001). Using the aerosol size distributions collected during the descent, aerosol optical depths were calculated (see Fig. 6) (Clarke et al. 1996). In this case the largest size the FSSP measured was 30-μm diameter and an index of refraction of 1.35 was used for the Mie calculations. Above the boundary layer inversion (>1400 m) optical depths show little change with height indicating aerosol extinction to be small at this altitude in the free troposphere. Below 1200 m, the aerosol optical depth increases rapidly with decreasing altitude due to larger sea salt aerosol concentrations. In general, optical depths calculated from size distributions measured by the FSSP show good agreement with manual sun photometer measurements. However, manual sun photometer measurements indicated substantially larger optical depths than those calculated from measured aerosol size distributions made near the surface possibly due to the misalignment issues discussed above. The FSSP was newly calibrated and considered to be working correctly during this experiment, but valid concerns have been raised about the validity of FSSP size distribution measurement (Reid et al. 2006).

Fig. 6.

Comparison of aerosol optical depth measurements from Microtops sun photometer and calculations based on FSSP size distribution measurements (mounted below the Seneca aircraft). Measurements were made on 19 Mar 2000 at 19.4°N, 157.8°W.

Fig. 6.

Comparison of aerosol optical depth measurements from Microtops sun photometer and calculations based on FSSP size distribution measurements (mounted below the Seneca aircraft). Measurements were made on 19 Mar 2000 at 19.4°N, 157.8°W.

Figure 7 shows that the spectral dependence of aerosol optical depths measured near the surface and at 3800 m was essentially identical. The coarse mode sea salt aerosol in the boundary layer increased the optical depth but did not change the wavelength dependence of the aerosol optical depth. Spectrally flat optical depths are typical of coarse mode sea salt because the aerosols are much larger than the wavelengths at which the aerosol optical depth measurements were made.

Fig. 7.

Spectral dependence of aerosol optical depth at surface and above trade wind inversion on 19 Mar 2000 over the Moby buoy on a strong wind day dominated by sea salt.

Fig. 7.

Spectral dependence of aerosol optical depth at surface and above trade wind inversion on 19 Mar 2000 over the Moby buoy on a strong wind day dominated by sea salt.

On 26 January 2000, aircraft measurements were obtained in the Hawaiian volcano plume in the lee (Kona side) of the “Big Island” of Hawaii. An image of the volcanic aerosol plume, taken from the Seneca aircraft (Fig. 8), shows a strong lid on the volcanic aerosols. This capping is created by the trade wind inversion at the top of the marine boundary layer. On this flight, sun photometer measurements made during transects of the volcano plume were used to calculate aerosol column average mass concentrations (10–15 μg m−3 sulfate), which were compared to Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) dispersion model calculations (Hollingshead et al. 2003). In contrast to the sea salt case discussed above, the volcanic aerosol optical depth has a strong spectral dependence (Fig. 9), which is consistent with accumulation mode aerosols. Inversion of sun photometer measurements (Porter et al. 2002) as well as in situ optical particle counter and differential mobility analyzer measurements (Porter and Clarke 1997) indicate most of the volcano aerosol existed in the accumulation mode.

Fig. 8.

Volcano plume capped by trade wind inversion in the lee of the Big Island of Hawaii.

Fig. 8.

Volcano plume capped by trade wind inversion in the lee of the Big Island of Hawaii.

Fig. 9.

Spectral measurements of aerosol optical depths below and above the Hawaii volcano plume.

Fig. 9.

Spectral measurements of aerosol optical depths below and above the Hawaii volcano plume.

During late summer 2004, aircraft and ground-based measurements were carried out as part of the UAE2. During this experiment, a Cheyenne aircraft was flown over the Mobile Atmospheric Aerosol and Radiation Characterization Observatory (MAARCO) supersite as well as other locations over the desert and the Arabian Gulf. Sun photometer measurements carried out on 18 September 2004 (over the MAARCO site at 24.7°N, 54.667°E) are shown in Fig. 10. Aerosol optical depths (at 500 nm) increased from ∼0.04, at 5000-m height, to ∼0.43 near the surface. This relatively large aerosol loading is typical of this Middle Eastern region due to pollution and dust production. Compared to measurements from the Seneca taken over Hawaii (Fig. 5), the fraction of reliable measurements (solid circles) made from the Cheyenne over the UAE (Fig. 10) was larger due to the relatively calmer conditions during this flight. Wind speeds were below 4 m s−1 with early morning convection just beginning.

Using the sun photometer measurements shown in Fig. 10, the vertical profile of aerosol optical depths were fit with a fifth-order polynomial fit as well as straight line fits to the average at each height (Fig. 11). Although these measurements appear reasonably good, significant variability exists about the fit. The difference between the measurements and the model is in part due to natural aerosol variability but may also be due to pointing errors. Based on the polynomial fits, aerosol extinction coefficients were calculated by taking differential vertical aerosol optical depths every 100 m (Clarke et al. 1996). Figure 12 shows the vertical distribution of aerosol extinction values. The aerosol extinction coefficients at different wavelengths all show a similar decrease with height decreasing by nearly a factor of 10 at 2000-m height.

Fig. 11.

Vertical profiles of aerosol optical depths (UAE experiment) and fifth-order polynomial fit for 440-, 500-, 675-, and 870-nm wavelengths.

Fig. 11.

Vertical profiles of aerosol optical depths (UAE experiment) and fifth-order polynomial fit for 440-, 500-, 675-, and 870-nm wavelengths.

Fig. 12.

Vertical profiles of aerosol extinction coefficients derived from the polynomial fits shown in Fig. 11 (for UAE2 experiment).

Fig. 12.

Vertical profiles of aerosol extinction coefficients derived from the polynomial fits shown in Fig. 11 (for UAE2 experiment).

It is interesting to point out that deriving aerosol extinction coefficients from differential aerosol optical depth measurements does not depend on accurate calibration of the sun photometer. This can be shown in the following way. Beginning with the Beer’s law I = Ioe(−τM), we can also write V = Voe(−τM) if the instrument detector is linear (which photodiodes are). Here V and Vo are the instrument voltages for direct sun light measured at the bottom and top of the atmosphere, I and Io are the direct light at the bottom and top of the atmosphere, τ is the total optical depth (aerosol plus gas, scatter plus absorption) in the atmosphere, and M is the airmass {[1/ cos(solar zenith angle)]} (Porter et al. 2001). Solving for the total optical depth, we obtain τ = (1/M) ln[(V/Vo)]. Next the extinction coefficient is defined as σ = (ΔτH), where Δτ is the change in total optical depth between two layers and ΔH is the change in height between the same two layers. Substituting τ into this last equation we see that the Vo term (the extraterrestrial constant required for calibration) drops out and the extinction coefficient calculation, σ = (1/ΔH M) ln[(V1/V2)], does not require calibration of the sun photometer. Here V1 and V2 are the voltages measured by the sun photometer at each height. The aerosol extinction coefficient (σa) is then obtained by subtracting the molecular extinction from the total extinction (σa = σ − σm) at that wavelength. It has also been assumed that the air mass (M) is the same for both measurements. This requires that the sun photometer measurements be taken close together in time and that they be taken 1.5 h after sunrise or ∼1.5 h before sunset as the air mass changes rapidly at those times. Therefore, it is possible to obtain vertically averaged extinction values with an uncalibrated sun photometer system.

While the overall shape of the extinction profile shown in Fig. 12 is reasonable, the spectral dependence of the aerosol extinction at different heights appears unreliable as they cross over each other. Such variability with wavelength may not be correct and is probably due to sun photometer misalignment during the measurements. The shape of the extinction curve is critically dependent on which sun photometer values are classified as reliable in rejecting the larger optical depths. In addition, the results depend on the vertical resolution of this analysis. In this calculation, extinction coefficients were calculated every 100-m altitude from the polynomial curve fit. If the aerosol optical depths had been differenced over a larger vertical distance, then the resulting values would have been more reliable but with poorer vertical resolution.

The spectral dependence of aerosol optical depths measured from the aircraft at 200- and 5000-m height (over the MAARCO site) are shown in Fig. 13. Aerosol optical depth measurements made from the surface AERONET sun photometer (at the same time) are also shown in the Fig. 13. To compare the aircraft and surface optical depth measurements, the additional optical depth, below the aircraft, was calculated from the extinction values in lowest height in Fig. 12. With this addition, the optical depth values measured from the aircraft are in good agreement with the AERONET with the exception of the 380-nm channel. Following the UAE2 experiment it was discovered that the 380-nm channel was malfunctioning and could not even be used to carry out a Langley plot calibration from measurements carried out at the Mauna Loa Observatory. Therefore, the 380-nm channel was considered to be unreliable for the UAE experiment. It is important to mention that the AERONET and Microtops calibration was based on independent Langley plot calibrations performed at the Mauna Loa Observatory. The good agreement suggests both calibrations were accurate.

Fig. 13.

Spectral aerosol optical depth at 200- and 5000-m altitude. Aerosol optical depth extrapolated from 200-m height to surface is shown along with surface Cimel measurements.

Fig. 13.

Spectral aerosol optical depth at 200- and 5000-m altitude. Aerosol optical depth extrapolated from 200-m height to surface is shown along with surface Cimel measurements.

5. Sun photometer camera system

An approach for collecting and processing manual sun photometer measurements from small aircraft was described above. Although the approach can provide quantitative optical depth measurements from aircraft, it requires a large number of measurements at each height of which many will be rejected. If only a few measurements remain after the screening process then the final results can be questionable. To address this problem we have carried out tests with a small camera mounted on the side of the Microtops sun photometer. The goal was to use the camera to image the sun as the sun photometer measurement is made. By calculating the position of the sun in the image, it should be possible to determine if the sun photometer was correctly pointed at the sun when the sun photometer measurement was made.

Figure 14 shows the configuration of the camera–Microtops system that we tested. To protect the camera, a neutral density absorbing filter was placed in front of the camera and the camera automatic gain control was turned off. Using computer software, camera images were collected before and after the Microtops measurements were triggered. The images were then processed to determine the xy position of the sun in both images, and the average position of the sun during the sun photometer measurement was then obtained. For these measurements the Microtops sun photometer was triggered using a serial port connection and the Microtops sampling was set to a minimum value. Figures 15 and 16 show the result of a test carried out with the system while making sun photometer measurements on the ground. Figure 15 shows the xy position of the sun for each measurement, while the z axis is the voltage measured by one of the Microtops channels. A clear maximum in Microtops voltage values occurs when the sun is in the correct xy position in the image. Figure 16 shows the same camera–Microtops measurements now plotted as a function of pixel distance, of the sun xy position, away from the correct xy position for proper sun alignment. The reasonably tight correlation suggests that the approach should work well in real measurements. Using a tilt meter, it was determined that the ∼2.5° region with acceptably constant values near the top corresponds to 0.34°. This results in an acceptance angle of ∼0.7° (full angle), which will produce good sun photometer measurements. This assessment is similar to the tests carried out by Massen (2005).

Fig. 14.

Microtops and webcam camera.

Fig. 14.

Microtops and webcam camera.

Fig. 15.

Sun photometer voltage value plotted vs camera pixel position. In this figure the sun position in the image is given by the camera x and y pixel position. The z axis is the sun photometer measurement (voltage value) made at the same time. The peak z values correspond to when the sun photometer is correctly pointed at the sun. The circle of lower z values is when the sun was circled around the outer circle of the Microtops pointing window.

Fig. 15.

Sun photometer voltage value plotted vs camera pixel position. In this figure the sun position in the image is given by the camera x and y pixel position. The z axis is the sun photometer measurement (voltage value) made at the same time. The peak z values correspond to when the sun photometer is correctly pointed at the sun. The circle of lower z values is when the sun was circled around the outer circle of the Microtops pointing window.

Fig. 16.

Sun photometer voltage value plotted vs camera pixel distance from correct alignment pixel position. The sun photometer is correctly aligned when the camera pixel distance is zero. The flat region near the top, where the sun photometer voltages are approximately constant is approximately 2.5 pixels wide (half angle), each camera pixel is ∼0.135° so that the half angle acceptance angle for good sun photometer measurements is ∼0.34° (or ∼0.7° full angle).

Fig. 16.

Sun photometer voltage value plotted vs camera pixel distance from correct alignment pixel position. The sun photometer is correctly aligned when the camera pixel distance is zero. The flat region near the top, where the sun photometer voltages are approximately constant is approximately 2.5 pixels wide (half angle), each camera pixel is ∼0.135° so that the half angle acceptance angle for good sun photometer measurements is ∼0.34° (or ∼0.7° full angle).

To test the new Storm Prediction Center (SPC) system a short aircraft flight was carried out just north of Oahu. Measurements were collected at several heights, and the position of the sun in the images was used to select measurements that were expected to correctly be pointed at the sun. Unfortunately, it was found that the sun being in the correct xy pixel position did not guarantee that those Microtops measurements were the lowest optical depths and therefore likely to be correctly aligned. Various tests in image analysis were carried out (varying acceptable range values, using only first or last images, etc.) all showing the approach was not working as expected. After tests with the Microtops and discussions with Solar Light, Inc., it was determined that the Microtops multiplexer takes ∼1.1 s to scan through each of the five channels once a measurement has been initiated (for the Microtops set to the minimum number of samples). Apparently the motion of the aircraft is sufficient to cause significant misalignment between the time required to scan over the five Microtops channels (∼1.1 s). Since each image takes only 10 ms to collect, it is likely that the misalignment occurs when the Microtops is scanning through its wavelengths. This interpretation of the problem is consistent with our past experience with the Microtops sun photometer. For both ship measurements (Porter et al. 2001) and the aircraft measurements (described above), it was found that even though aerosol optical depths in one channel were determined to be accurate (pointed at the sun) it did not guarantee that those made by the other channels would be. This required each channel to be processed separately to select the minimum values for both the ship (Porter et al. 2001) and the aircraft measurements (see Figs. 5 and 10). Although the camera approach tested here failed, it is expected that if this approach were used with a different sun photometer system, which collected all the channels simultaneously (such as a spectrometer diode array), then it could provide a simple way to test the sun photometer alignment of each measurement.

6. Summary

Manual sun photometer measurements offer a relatively inexpensive and simple way to obtain vertical profiles of aerosol optical depths from light aircraft. This approach requires significant effort to obtain a large number of optical depth measurements of which only 5%–10% will be considered useful after the erroneous larger aerosol optical depths have been eliminated. Only if sufficient measurements remain after the necessary screening process can some amount of assurance of data quality be obtained. Although laborious, the approach routinely produces good quality measurements. To confirm the pointing accuracy of each Microtops measurement, a camera was added to the side of a Microtops sun photometer. Although ground-based measurements were promising, a test flight shows that the 1.1 s the Microtops takes to make its five-channel measurements was too long and allowed movement of the system after the camera measurement was made. Although the camera approach did not work well for the Microtops sun photometer, it is expected to work well for a system that collects all the measurements at the same time (such as a diode array).

Acknowledgments

This work was supported by NASA Grant 653828. NRL’s participation was funded by the Office of Naval Research Code 32 (N0001405WR20206).

REFERENCES

REFERENCES
Chin
,
M.
, and
Coauthors
,
2002
:
Tropospheric aerosol optical thickness from the GOCART model and comparisons with satellite and sun photometer measurements.
J. Atmos. Sci.
,
59
,
461
483
.
Clarke
,
A.
,
J.
Porter
,
F.
Valero
, and
P.
Pilewskie
,
1996
:
Vertical profiles, aerosol microphysics, and optical closure during the Atlantic Stratocumulus Transition Experiment: Measured and modeled column optical properties.
J. Geophys. Res.
,
101
,
4443
4454
.
Frouin
,
R.
,
B.
Holben
,
M.
Miller
,
C.
Pietras
,
J.
Porter
, and
K.
Voss
,
2001
:
Sun and sky radiance measurements and data analysis protocols. In Situ Aerosol Optical Thickness Collected by the SIMBIOS Program (1997–2000): Protocols, and Data QC and Analysis, G. Fargion et al., Eds., NASA Tech. Memo. 2001-209982, NASA Goddard Space Flight Center, Greenbelt, MD, 26–42
.
Harrison
,
L.
,
J. J.
Michalsky
, and
J.
Berndt
,
1994
:
Automated multifilter rotation shadowband radiometer: An instrument for optical depth and radiation measurements.
Appl. Opt.
,
33
,
5118
5125
.
Holben
,
B. N.
, and
Coauthors
,
1998
:
A federated instrument network and data archive for aerosol characterization.
Remote Sens. Environ.
,
66
,
1
16
.
Hollingshead
,
A. T.
,
S.
Businger
,
R.
Draxler
,
J. N.
Porter
, and
D.
Stevens
,
2003
:
Dispersion modeling of the Kilauea Plume.
Bound.-Layer Meteor.
,
108
,
121
144
.
Ichoku
,
C.
, and
Coauthors
,
2002
:
Analysis of the performance characteristics of the five-channel Microtops II sun photometer for measuring aerosol optical thickness and precipitable water vapor.
J. Geophys. Res.
,
107
.
4179, doi:10.1029/2001JD001302
.
Ignatov
,
A.
, and
L.
Stowe
,
2002
:
Aerosol retrievals from individual AVHRR channels. Part II: Quality control, probability distribution functions, information content, and consistency checks of retrievals.
J. Atmos. Sci.
,
59
,
335
362
.
Knobelspiesse
,
K. D.
,
C.
Pietras
, and
G. S.
Fargion
,
2003
:
Sun-pointing-error correction for sea deployment of the MICROTOPS II handheld sun photometer.
J. Atmos. Oceanic Technol.
,
20
,
767
771
.
Korotaev
,
G.
,
S. M.
Sakerin
,
A. M.
Ignatov
,
L. L.
Stowe
, and
E. P.
McClain
,
1993
:
Sun-photometer observations of aerosol optical thickness over the North Atlantic from a Soviet research vessel for validation of satellite measurements.
J. Atmos. Oceanic Technol.
,
10
,
725
735
.
Maring
,
H.
,
D. L.
Savoie
,
M. A.
Izaguirre
, and
L.
Custals
,
2003
:
Vertical distribution of dust and sea-salt aerosols over Puerto Rico during PRIDE measured from a light aircraft.
J. Geophys. Res.
,
108
.
8587, doi:10.1029/2002JD002544
.
Massen
,
F.
,
cited
.
2005
:
Influence of intentionally mis-pointing on Microtops II readings. [Available online at http://www.restena.lu/meteo_lcd/papers/index.html.]
.
Matsumoto
,
T.
,
P. B.
Russell
,
C.
Mina
,
W.
Van Ark
, and
V.
Banta
,
1987
:
Airborne tracking sunphotometer.
J. Atmos. Oceanic Technol.
,
4
,
336
339
.
Miller
,
M. A.
,
M. J.
Bartholomew
, and
R. M.
Reynolds
,
2004
:
The accuracy of marine shadow-band sun photometer measurements of aerosol optical thickness and angstrom exponent.
J. Atmos. Oceanic Technol.
,
21
,
397
410
.
Morys
,
M. F.
,
F. M.
Mims
III
, and
S. E.
Anderson
,
cited
.
1996
:
Design, calibration and performance of MICROTOPS II hand-held ozonometer. [Available online at http://www.solarlight.com/download/mtops.pdf.]
.
Penner
,
J. E.
, and
Coauthors
,
2002
:
A comparison of model- and satellite-derived aerosol optical depth and reflectivity.
J. Atmos. Sci.
,
59
,
441
460
.
Porter
,
J. N.
, and
A. D.
Clarke
,
1997
:
Aerosol size distribution models based on in situ measurements.
J. Geophys. Res.
,
102
,
6035
6046
.
Porter
,
J. N.
,
A. D.
Clarke
, and
B. R.
Lienert
,
2000
:
Aircraft/surface-derived aerosol optical properties near Hawaii for satellite validation. Proc. SPIE, 4150, 274–281
.
Porter
,
J. N.
,
M.
Miller
,
C.
Pietras
, and
C.
Motell
,
2001
:
Ship-based sun photometer measurements using Microtops sun photometers.
J. Atmos. Oceanic Technol.
,
18
,
765
774
.
Porter
,
J. N.
, and
Coauthors
,
2002
:
Sun photometer and lidar measurements of the plume from the Hawaii Kilauea volcano Pu’u O’o Vent: Aerosol flux and SO2 lifetime.
Geophys. Res. Lett.
,
29
.
1783, doi:10.1029/2002GL014744
.
Redemann
,
J.
,
Q.
Zhang
,
B.
Schmid
,
P. B.
Russell
,
J. M.
Livingston
,
H.
Jonsson
, and
L. A.
Remer
,
2006
:
Assessment of MODIS-derived visible and near-IR aerosol optical properties and their spatial variability in the presence of mineral dust.
Geophys. Res. Lett.
,
33
.
L18814, doi:10.1029/2006GL026626
.
Reid
,
J. S.
, and
Coauthors
,
2006
:
Reconciliation of coarse mode sea-salt aerosol particle size measurements and parameterizations at a subtropical ocean receptor site.
J. Geophys. Res.
,
111
.
D02202, doi:10.1029/2005JD006200
.
Reynolds
,
M. R.
,
M. A.
Miller
, and
M. J.
Bartholomew
,
2001
:
A fast-rotating spectral shadow-band radiometer for marine applications.
J. Atmos. Oceanic Technol.
,
18
,
200
214
.
Shaw
,
G. E.
,
1975
:
The vertical distribution of tropospheric aerosols at Barrow, Alaska.
Tellus
,
27
,
39
49
.
Shaw
,
G. E.
,
1983
:
Sun photometry.
Bull. Amer. Meteor. Soc.
,
64
,
4
10
.
Valero
,
F. P. J.
,
T. P.
Ackerman
, and
W. J. Y.
Gore
,
1989
:
The effects of the Arctic haze as determined from airborne radiometric measurements during AGASP II.
J. Atmos. Chem.
,
9
,
225
244
.
Volz
,
F. E.
,
1959
:
Photometer mit Selen-Photoelement zur spektralen Messung der Sonnenstrahlung und zur Bestimmung der Wellenlängenabhängigkeit der Dunsttrübung.
Arch. Meteor. Geophys. Bioklimatol.
,
B10
,
100
131
.

Footnotes

Corresponding author address: John N. Porter, 2525 Correa Road, HIG room 313, Honolulu, HI 96822. Email: johnport@hawaii.edu

* School of Ocean and Earth Science and Technology Contribution Number 6688.