REVIEW OF SCIENTIFIC INSTRUMENTS 85, 066104 (2014)
Note: Active optical detection of cloud from a balloon platform R. G. Harrison and K. A. Nicoll Department of Meteorology, University of Reading, Reading RG6 6BB, United Kingdom
(Received 22 November 2013; accepted 27 May 2014; published online 9 June 2014) A disposable backscatter instrument is described for optical detection of cloud in the atmosphere from a balloon-carried platform. It uses an ultra-bright light emitting diode (LED) illumination source with a photodiode detector. Scattering of the LED light by cloud droplets generates a small optical signal which is separated from background light fluctuations using a lock-in technique. The signal to noise obtained permits cloud detection using the scattered LED light, even in daytime. The response is interpreted in terms of the equivalent visual range within the cloud. The device is lightweight (150 g) and low power (∼30 mA), for use alongside a conventional meteorological radiosonde. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4882318] Clouds are important in controlling the radiation received and emitted at the earth’s surface, in both the infra-red and visible regions of the electromagnetic spectrum. The relative contributions of the infra-red and visible radiation to the energy balance depend on both the cloud height and the droplet properties which determine the scattering and infra-red emissivity near to the cloud edge.1 For extensive horizontal layer clouds, the cloud-air boundary also represents an electrical conductivity transition, where droplet electrification can be observed due to vertical current flow. Accurate retrieval of the cloud-air transition is important in verifying the location and magnitude of cloud edge electrification, but uncertainties associated with capacitance radiosonde relative humidity (RH) sensors, such as their slow time response,2 limits the resolution with which sharp cloud to air transitions can be identified. Optical detection of cloud from meteorological radiosondes can increase the vertical resolution over the standard thermodynamic methods. It can also help to detect thin clouds which would otherwise be poorly identified due to the time constants of capacitance sensors. For example, “passive” sensing3 of scattered sunlight using a balloon-carried photodiode sensor has shown improved cloud boundary detection over that inferred from capacitance humidity measurements, and is sufficiently inexpensive to be disposable. In situ optical cloud sensing is extended here beyond the passive method to allow cloud detection sensing in low light and nocturnal conditions by carrying a local “active” source of illumination. Previous active approaches have been bulky4 or only suitable for nocturnal use.5 Random balloon motions preclude laser sources due to safety considerations, but ultrabright light emitting diodes (LEDs) now permit a lightweight, low power, and inexpensive instrument for use even in daytime to complement the passive cloud detection technique. The instrument concept is given in Figure 1(a). Light entering the photodiode detector from all sources is amplified and provided as an output as in Ref. 3, but a modulated ultrabright LED light source is also included. LED light scattered by cloud droplets which reaches the photodiode is extracted from background light by the combination of high pass filtering and synchronous detection. After a low pass filter, the LED light signal is substantially amplified for 16 bit data
0034-6748/2014/85(6)/066104/3/$30.00
acquisition using a “PANDORA” (Programmable ANalogue and Digital Operational Radiosonde Accessory).6 This interface also provides bipolar power rails, and embeds the measurements within the radiosonde’s ultrahigh frequency (UHF) data telemetry to a ground station. Figure 1(b) provides an overview of the signal conditioning employed using analogue opamp circuitry. The same photodiode (type VTB8440B, peak spectral response at 580 nm) used in (3) provides the detector (D1), alongside two ultrabright (maximum output power 182 mW, luminous intensity 100 cd) yellow (590 nm) LEDs (Tru Opto type OS5YKA5111P) as sources. The LEDs (D2 and D3) are driven in series to minimize power conversion losses from a square wave oscillator (modulating frequency ∼1 kHz), and are mounted, together with the photodiode, in parallel collimators machined from black delrin polymer (polyoxymethylene) to ensure only backscattered LED light can enter the photodiode. If there is no backscattered LED light, only scattered sunlight will be received by the photodiode (see also Figure 1(a)). The D1 photocurrent is amplified by a single opamp stage (feedback resistor R1 = 20 k and capacitor C1 = 1 nF), with the background light level signal (V1 ) used directly for passive sensing of cloud, as previously.3 This signal is also high-pass filtered, amplified (×100) at non-inverting stage G1, and then synchronously demodulated to reject variations not originating from the modulated light source. After low pass smoothing, the demodulated signal is further amplified (×310) by a non-inverting stage, G2. Finally, to ensure a positive output voltage is maintained with or without cloud present, a 2.5 V offset is added to the amplified demodulated signal. This offset voltage compensates for offset voltage errors and phase delay (measured as 11◦ at the phase detector for a 1 kHz signal in the laboratory), and keeps the output signal V2 within the unipolar measurement range of the PANDORA interface, despite drift from the substantial temperature changes encountered during a balloon flight. The operating principle and sensitivity of the instrument was tested in a laboratory experiment in which small water droplets (size range 1 to 10 μm) were generated by ultrasonic atomization. A small electric fan was used to draw the water
85, 066104-1
© 2014 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 202.177.173.189 On: Tue, 01 Jul 2014 05:23:18
066104-2
R. G. Harrison and K. A. Nicoll
Rev. Sci. Instrum. 85, 066104 (2014)
FIG. 1. (a) Concept of the active cloud detector employing a collimated ultrabright LED source (D2, D3) and photodiode (D1) to detect backscattered radiation. (b) Block diagram of the analogue signal conditioning circuitry, showing photocurrent amplifier (output V1), high pass (HP), and low pass (LP) filters, phase-locked demodulator (S1) and gain stages (G1, G2). oscillator (OSC) represents the (kHz) modulating signal. The −1 gain stage provides signal inversion, and the final summing stage adds a 2.5 V offset to ensure the signal voltage, V2 is always positive.
droplets into the optical system, allowing pulses of droplets to be generated, forming transiently, a range of droplet concentrations. Two identical measurement systems B (for backscatter) and T (for transmission) were used, both sampling the same cloud of water droplets. System B was configured as in Figure 1 to measure the LED light scattered from the water droplets, but for T, the LED illumination source was displaced to allow the light to be transmitted through the water droplets. This additional measurement of the transmitted light allows the optical extinction variations during the pulse to be calculated, simultaneously with the backscatter measurement. Extinction and backscatter are closely related, but the functional form shows some sensitivity to the size and shape of the droplets or particles concerned.7 (The relationship between extinction and backscatter is also a key parameter in lidar retrievals of atmospheric properties, in general unknown.8 ) Extinction can be related to the visual range, a commonly-made meteorological measurement. For example, the laser detection of cloud base9 is associated with a reduction of visual range at the cloud base to about 100 m. The extinction ξ typically varies from 0.01 km−1 in clear air to 10 km−1 or greater in fog10 and 50 km−1 or greater in mixed phase clouds.11 The associated visual range X is X=−
ln ε , ξ
(1)
where ε is the visual contrast factor, usually12 ε = 0.05. In the laboratory system T, with a received signal I0 before the droplet pulse is introduced, the time variation in the transmitted signal I(t) during the droplet pulse is given by I (t) = I0 exp[−ξ (t)x],
(2)
where x is the path length (36 cm) over which the attenuation occurs and ξ (t) is the (time-varying) extinction co-
FIG. 2. Derived optical extinction ξ plotted against the measured backscattered voltage Vb , normalized by the current supplied to the LED source (0.3 ± 0.1 mA). The right-hand axis shows the equivalent visual range from Eq. (1). The measurements were binned into equal intervals to determine the mean value in the interval, and the error bar represents one standard error; a linear fit ξ = a(Vb / iLED ) + b has been added for those visual ranges typical of in-cloud conditions, i.e., between 10 and 100 m, giving a = (11.88 ± 0.67) km−1 (mV mA−1 )−1 and b = (−118 ± 12.5) km−1 .
efficient. From measurements of a series of droplet pulses, Figure 2 shows the extinctions derived from the transmitted voltage using Eq. (2) plotted against the simultaneous backscatter detector voltages Vb after demodulation. The extinction is also used to determine the visual range using Eq. (1), which is shown on an additional vertical axis. In the laboratory experiment with small geometry and substantial droplet concentrations (measured as typically between 500 and 1000 cm−3 using an ASPECT particle analyser), a reduced LED current (0.3 mA) was used to diminish the signal. To detect atmospheric water clouds, the brightness of the LED source in the balloon instrument is increased by using a larger driving current (20 mA) to generate more backscatter. The received backscatter voltage in Figure 2 has therefore been normalized by the LED current used in the laboratory experiment, to allow the expected response in backscatter voltage to be scaled. (A subsidiary experiment showed that the brightness of the LED varied linearly with driving current.) For use with a radiosonde, the electronic circuitry was constructed on double-sided printed circuit board, with the stages linearly arranged to physically separate input and output. This was mounted with the PANDORA interface in a shielded box (130 × 70 × 45 mm), attached to the side of a Vaisala RS92 meteorological radiosonde. Combined with the interface, the total mass was 150 g and the current consumption was 30 mA at 9 V. No additional power source was needed beyond the standard RS92 radiosonde battery of six alkaline AA cells. Figure 3 shows measurements made by the instrument during a daylight balloon sounding through a cloud layer of liquid water droplets (stratocumulus) about 400 m thick, with cloud base at 1250 to 1270 m. The standard meteorological measurements of temperature and relative humidity were obtained from the RS92 radiosonde. These evidently only provide a coarse indication of the cloud position as the threshold humidity at which cloud forms is poorly defined, and the cloud boundary determination is also limited by the time re-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 202.177.173.189 On: Tue, 01 Jul 2014 05:23:18
066104-3
R. G. Harrison and K. A. Nicoll
Rev. Sci. Instrum. 85, 066104 (2014)
FIG. 3. Vertical sounding made from Reading in daylight conditions through a cloud layer showing (a) thermodynamic quantities of Relative Humidity and temperature measured by a Vaisala RS92 radiosonde and (b) simultaneous passive (solar radiation) and active sensing (visual range) of the cloud. (The visual range was derived from the response in Figure 2.)
sponse of the relative humidity sensor. A temperature inversion around 1600 m provides a further indication of the cloud top region. The solar radiation from the photodiode, Figure 3(b) (V1 from Figure 1(b)) shows the characteristic variation of increasing solar radiation as the instrument ascends through the cloud, followed, on its emergence from the cloud into sunlight, by increased variability. This transition, together with the increase at the cloud base, indicates the extent of the cloud, as discussed previously.3 Figure 3(b) also shows the demodulated output obtained from the LED light source (V2 from Figure 1(b)), which increases when light is scattered back into the photodiode from cloud droplets or other hydrometeors present. This has been calibrated to visual range, using X=−
ln ε a(Vb / iLED ) + b
similar heights to the indications from the thermodynamic measurements. This device can operate in daylight or at night. Even in daylight, the combination of the active and passive signals extracted from the same detector shows greatly improved vertical resolution of cloud boundaries over the standard meteorological determination using thermodynamic radiosonde sensors. This work was supported by the Natural Environment Research Council, UK, grant reference NE/H002081/1 (More Operational Radiosonde Sensors, MORSE) and by a Leverhulme Trust Early Career Fellowship for KAN. A. G. Lomas developed the mechanical and electrical arrangements; R. Wilson and I. Read assisted with construction of prototypes and balloon launches. Biral (www.biral.com) provided the ASPECT particle sizer.
(3)
with a and b from the in-cloud linear relationship derived in Figure 2, scaled to the increased LED current of 20 mA. The error bars in Figure 2 indicate an uncertainty in visual range of ±30 m, but the results for cloud detection purposes are not inconsistent with the usual criterion for detecting cloud base9 of 100 m visual range. In terms of absolute voltage response, the backscatter response to cloud was detectable at about ∼10 mV. Allowing for the total signal gain (G1 × G2), the ∼10 mV minimum detectable demodulated response against the noise floor corresponds to about 0.3 μV at the output of U1, in a background signal level of ∼0.3 V. This amounts to extracting the signal from the background at one part in a million, as can be expected for the lockin technique.13 The cloud detector shows a sharp response at the upper and lower cloud boundaries, qualitatively at
1 I.
Koren, L. A. Remer, Y. J. Kaufman et al., Geophys. Res. Lett. 34, L08805, doi:10.1029/2007GL029253 (2007). 2 L. M. Miloshevich, A. Pakkunen, H. Vomel, and S. J. Oltmans, J. Atmos. Oceanic Technol. 21, 1305–1327 (2004). 3 K. A. Nicoll and R. G. Harrison, Rev. Sci. Instrum. 83, 025111 (2012). 4 J. M. Rosen and N. T. Kjome, Appl. Opt. 30(12), 1552–1561 (1991). 5 See http://www.iac.ethz.ch/groups/peter/research/Balloon_soundings/ COBALD_sensor for details of the COBALD aerosol sensor. 6 R. G. Harrison, K. A. Nicoll, and A. G. Lomas, Rev. Sci. Instrum. 83, 036106 (2012). 7 R. W. Fenn, Appl. Opt. 5, 293–295 (1966). 8 V. A. Kovalev and W. E. Eichinger, Elastic Lidar: Theory, Practice, and Analysis Methods (Wiley, 2004). 9 Ceilometer CT25 User’s Guide, CT25K-U059en-2.1 (Vaisala, 1999). 10 Handbook of Meteorological Instruments: Measurement of Visibility and Cloud Height (Her Majesty’s Stationery Office, 1982), Vol. 7. 11 A. Korolev, A. Shashkov, H. Barker, US Department of Energy Report DOE/SC-ARM-TR-105, ARM Climate Research Facility (2012). 12 H. Koschmieder, Beitr. Phys. Freien Atmos. 12, 171 (1924). 13 P. Horowitz and W. Hill, The Art of Electronics (Cambridge University Press, Cambridge, 1989).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 202.177.173.189 On: Tue, 01 Jul 2014 05:23:18
Review of Scientific Instruments is copyrighted by the American Institute of Physics (AIP). Redistribution of journal material is subject to the AIP online journal license and/or AIP copyright. For more information, see http://ojps.aip.org/rsio/rsicr.jsp
Note: Active optical detection of cloud from a balloon platform R. G. Harrison and K. A. Nicoll Department of Meteorology, University of Reading, Reading RG6 6BB, United Kingdom
(Received 22 November 2013; accepted 27 May 2014; published online 9 June 2014) A disposable backscatter instrument is described for optical detection of cloud in the atmosphere from a balloon-carried platform. It uses an ultra-bright light emitting diode (LED) illumination source with a photodiode detector. Scattering of the LED light by cloud droplets generates a small optical signal which is separated from background light fluctuations using a lock-in technique. The signal to noise obtained permits cloud detection using the scattered LED light, even in daytime. The response is interpreted in terms of the equivalent visual range within the cloud. The device is lightweight (150 g) and low power (∼30 mA), for use alongside a conventional meteorological radiosonde. © 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4882318] Clouds are important in controlling the radiation received and emitted at the earth’s surface, in both the infra-red and visible regions of the electromagnetic spectrum. The relative contributions of the infra-red and visible radiation to the energy balance depend on both the cloud height and the droplet properties which determine the scattering and infra-red emissivity near to the cloud edge.1 For extensive horizontal layer clouds, the cloud-air boundary also represents an electrical conductivity transition, where droplet electrification can be observed due to vertical current flow. Accurate retrieval of the cloud-air transition is important in verifying the location and magnitude of cloud edge electrification, but uncertainties associated with capacitance radiosonde relative humidity (RH) sensors, such as their slow time response,2 limits the resolution with which sharp cloud to air transitions can be identified. Optical detection of cloud from meteorological radiosondes can increase the vertical resolution over the standard thermodynamic methods. It can also help to detect thin clouds which would otherwise be poorly identified due to the time constants of capacitance sensors. For example, “passive” sensing3 of scattered sunlight using a balloon-carried photodiode sensor has shown improved cloud boundary detection over that inferred from capacitance humidity measurements, and is sufficiently inexpensive to be disposable. In situ optical cloud sensing is extended here beyond the passive method to allow cloud detection sensing in low light and nocturnal conditions by carrying a local “active” source of illumination. Previous active approaches have been bulky4 or only suitable for nocturnal use.5 Random balloon motions preclude laser sources due to safety considerations, but ultrabright light emitting diodes (LEDs) now permit a lightweight, low power, and inexpensive instrument for use even in daytime to complement the passive cloud detection technique. The instrument concept is given in Figure 1(a). Light entering the photodiode detector from all sources is amplified and provided as an output as in Ref. 3, but a modulated ultrabright LED light source is also included. LED light scattered by cloud droplets which reaches the photodiode is extracted from background light by the combination of high pass filtering and synchronous detection. After a low pass filter, the LED light signal is substantially amplified for 16 bit data
0034-6748/2014/85(6)/066104/3/$30.00
acquisition using a “PANDORA” (Programmable ANalogue and Digital Operational Radiosonde Accessory).6 This interface also provides bipolar power rails, and embeds the measurements within the radiosonde’s ultrahigh frequency (UHF) data telemetry to a ground station. Figure 1(b) provides an overview of the signal conditioning employed using analogue opamp circuitry. The same photodiode (type VTB8440B, peak spectral response at 580 nm) used in (3) provides the detector (D1), alongside two ultrabright (maximum output power 182 mW, luminous intensity 100 cd) yellow (590 nm) LEDs (Tru Opto type OS5YKA5111P) as sources. The LEDs (D2 and D3) are driven in series to minimize power conversion losses from a square wave oscillator (modulating frequency ∼1 kHz), and are mounted, together with the photodiode, in parallel collimators machined from black delrin polymer (polyoxymethylene) to ensure only backscattered LED light can enter the photodiode. If there is no backscattered LED light, only scattered sunlight will be received by the photodiode (see also Figure 1(a)). The D1 photocurrent is amplified by a single opamp stage (feedback resistor R1 = 20 k and capacitor C1 = 1 nF), with the background light level signal (V1 ) used directly for passive sensing of cloud, as previously.3 This signal is also high-pass filtered, amplified (×100) at non-inverting stage G1, and then synchronously demodulated to reject variations not originating from the modulated light source. After low pass smoothing, the demodulated signal is further amplified (×310) by a non-inverting stage, G2. Finally, to ensure a positive output voltage is maintained with or without cloud present, a 2.5 V offset is added to the amplified demodulated signal. This offset voltage compensates for offset voltage errors and phase delay (measured as 11◦ at the phase detector for a 1 kHz signal in the laboratory), and keeps the output signal V2 within the unipolar measurement range of the PANDORA interface, despite drift from the substantial temperature changes encountered during a balloon flight. The operating principle and sensitivity of the instrument was tested in a laboratory experiment in which small water droplets (size range 1 to 10 μm) were generated by ultrasonic atomization. A small electric fan was used to draw the water
85, 066104-1
© 2014 AIP Publishing LLC
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 202.177.173.189 On: Tue, 01 Jul 2014 05:23:18
066104-2
R. G. Harrison and K. A. Nicoll
Rev. Sci. Instrum. 85, 066104 (2014)
FIG. 1. (a) Concept of the active cloud detector employing a collimated ultrabright LED source (D2, D3) and photodiode (D1) to detect backscattered radiation. (b) Block diagram of the analogue signal conditioning circuitry, showing photocurrent amplifier (output V1), high pass (HP), and low pass (LP) filters, phase-locked demodulator (S1) and gain stages (G1, G2). oscillator (OSC) represents the (kHz) modulating signal. The −1 gain stage provides signal inversion, and the final summing stage adds a 2.5 V offset to ensure the signal voltage, V2 is always positive.
droplets into the optical system, allowing pulses of droplets to be generated, forming transiently, a range of droplet concentrations. Two identical measurement systems B (for backscatter) and T (for transmission) were used, both sampling the same cloud of water droplets. System B was configured as in Figure 1 to measure the LED light scattered from the water droplets, but for T, the LED illumination source was displaced to allow the light to be transmitted through the water droplets. This additional measurement of the transmitted light allows the optical extinction variations during the pulse to be calculated, simultaneously with the backscatter measurement. Extinction and backscatter are closely related, but the functional form shows some sensitivity to the size and shape of the droplets or particles concerned.7 (The relationship between extinction and backscatter is also a key parameter in lidar retrievals of atmospheric properties, in general unknown.8 ) Extinction can be related to the visual range, a commonly-made meteorological measurement. For example, the laser detection of cloud base9 is associated with a reduction of visual range at the cloud base to about 100 m. The extinction ξ typically varies from 0.01 km−1 in clear air to 10 km−1 or greater in fog10 and 50 km−1 or greater in mixed phase clouds.11 The associated visual range X is X=−
ln ε , ξ
(1)
where ε is the visual contrast factor, usually12 ε = 0.05. In the laboratory system T, with a received signal I0 before the droplet pulse is introduced, the time variation in the transmitted signal I(t) during the droplet pulse is given by I (t) = I0 exp[−ξ (t)x],
(2)
where x is the path length (36 cm) over which the attenuation occurs and ξ (t) is the (time-varying) extinction co-
FIG. 2. Derived optical extinction ξ plotted against the measured backscattered voltage Vb , normalized by the current supplied to the LED source (0.3 ± 0.1 mA). The right-hand axis shows the equivalent visual range from Eq. (1). The measurements were binned into equal intervals to determine the mean value in the interval, and the error bar represents one standard error; a linear fit ξ = a(Vb / iLED ) + b has been added for those visual ranges typical of in-cloud conditions, i.e., between 10 and 100 m, giving a = (11.88 ± 0.67) km−1 (mV mA−1 )−1 and b = (−118 ± 12.5) km−1 .
efficient. From measurements of a series of droplet pulses, Figure 2 shows the extinctions derived from the transmitted voltage using Eq. (2) plotted against the simultaneous backscatter detector voltages Vb after demodulation. The extinction is also used to determine the visual range using Eq. (1), which is shown on an additional vertical axis. In the laboratory experiment with small geometry and substantial droplet concentrations (measured as typically between 500 and 1000 cm−3 using an ASPECT particle analyser), a reduced LED current (0.3 mA) was used to diminish the signal. To detect atmospheric water clouds, the brightness of the LED source in the balloon instrument is increased by using a larger driving current (20 mA) to generate more backscatter. The received backscatter voltage in Figure 2 has therefore been normalized by the LED current used in the laboratory experiment, to allow the expected response in backscatter voltage to be scaled. (A subsidiary experiment showed that the brightness of the LED varied linearly with driving current.) For use with a radiosonde, the electronic circuitry was constructed on double-sided printed circuit board, with the stages linearly arranged to physically separate input and output. This was mounted with the PANDORA interface in a shielded box (130 × 70 × 45 mm), attached to the side of a Vaisala RS92 meteorological radiosonde. Combined with the interface, the total mass was 150 g and the current consumption was 30 mA at 9 V. No additional power source was needed beyond the standard RS92 radiosonde battery of six alkaline AA cells. Figure 3 shows measurements made by the instrument during a daylight balloon sounding through a cloud layer of liquid water droplets (stratocumulus) about 400 m thick, with cloud base at 1250 to 1270 m. The standard meteorological measurements of temperature and relative humidity were obtained from the RS92 radiosonde. These evidently only provide a coarse indication of the cloud position as the threshold humidity at which cloud forms is poorly defined, and the cloud boundary determination is also limited by the time re-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 202.177.173.189 On: Tue, 01 Jul 2014 05:23:18
066104-3
R. G. Harrison and K. A. Nicoll
Rev. Sci. Instrum. 85, 066104 (2014)
FIG. 3. Vertical sounding made from Reading in daylight conditions through a cloud layer showing (a) thermodynamic quantities of Relative Humidity and temperature measured by a Vaisala RS92 radiosonde and (b) simultaneous passive (solar radiation) and active sensing (visual range) of the cloud. (The visual range was derived from the response in Figure 2.)
sponse of the relative humidity sensor. A temperature inversion around 1600 m provides a further indication of the cloud top region. The solar radiation from the photodiode, Figure 3(b) (V1 from Figure 1(b)) shows the characteristic variation of increasing solar radiation as the instrument ascends through the cloud, followed, on its emergence from the cloud into sunlight, by increased variability. This transition, together with the increase at the cloud base, indicates the extent of the cloud, as discussed previously.3 Figure 3(b) also shows the demodulated output obtained from the LED light source (V2 from Figure 1(b)), which increases when light is scattered back into the photodiode from cloud droplets or other hydrometeors present. This has been calibrated to visual range, using X=−
ln ε a(Vb / iLED ) + b
similar heights to the indications from the thermodynamic measurements. This device can operate in daylight or at night. Even in daylight, the combination of the active and passive signals extracted from the same detector shows greatly improved vertical resolution of cloud boundaries over the standard meteorological determination using thermodynamic radiosonde sensors. This work was supported by the Natural Environment Research Council, UK, grant reference NE/H002081/1 (More Operational Radiosonde Sensors, MORSE) and by a Leverhulme Trust Early Career Fellowship for KAN. A. G. Lomas developed the mechanical and electrical arrangements; R. Wilson and I. Read assisted with construction of prototypes and balloon launches. Biral (www.biral.com) provided the ASPECT particle sizer.
(3)
with a and b from the in-cloud linear relationship derived in Figure 2, scaled to the increased LED current of 20 mA. The error bars in Figure 2 indicate an uncertainty in visual range of ±30 m, but the results for cloud detection purposes are not inconsistent with the usual criterion for detecting cloud base9 of 100 m visual range. In terms of absolute voltage response, the backscatter response to cloud was detectable at about ∼10 mV. Allowing for the total signal gain (G1 × G2), the ∼10 mV minimum detectable demodulated response against the noise floor corresponds to about 0.3 μV at the output of U1, in a background signal level of ∼0.3 V. This amounts to extracting the signal from the background at one part in a million, as can be expected for the lockin technique.13 The cloud detector shows a sharp response at the upper and lower cloud boundaries, qualitatively at
1 I.
Koren, L. A. Remer, Y. J. Kaufman et al., Geophys. Res. Lett. 34, L08805, doi:10.1029/2007GL029253 (2007). 2 L. M. Miloshevich, A. Pakkunen, H. Vomel, and S. J. Oltmans, J. Atmos. Oceanic Technol. 21, 1305–1327 (2004). 3 K. A. Nicoll and R. G. Harrison, Rev. Sci. Instrum. 83, 025111 (2012). 4 J. M. Rosen and N. T. Kjome, Appl. Opt. 30(12), 1552–1561 (1991). 5 See http://www.iac.ethz.ch/groups/peter/research/Balloon_soundings/ COBALD_sensor for details of the COBALD aerosol sensor. 6 R. G. Harrison, K. A. Nicoll, and A. G. Lomas, Rev. Sci. Instrum. 83, 036106 (2012). 7 R. W. Fenn, Appl. Opt. 5, 293–295 (1966). 8 V. A. Kovalev and W. E. Eichinger, Elastic Lidar: Theory, Practice, and Analysis Methods (Wiley, 2004). 9 Ceilometer CT25 User’s Guide, CT25K-U059en-2.1 (Vaisala, 1999). 10 Handbook of Meteorological Instruments: Measurement of Visibility and Cloud Height (Her Majesty’s Stationery Office, 1982), Vol. 7. 11 A. Korolev, A. Shashkov, H. Barker, US Department of Energy Report DOE/SC-ARM-TR-105, ARM Climate Research Facility (2012). 12 H. Koschmieder, Beitr. Phys. Freien Atmos. 12, 171 (1924). 13 P. Horowitz and W. Hill, The Art of Electronics (Cambridge University Press, Cambridge, 1989).
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitationnew.aip.org/termsconditions. Downloaded to IP: 202.177.173.189 On: Tue, 01 Jul 2014 05:23:18
Review of Scientific Instruments is copyrighted by the American Institute of Physics (AIP). Redistribution of journal material is subject to the AIP online journal license and/or AIP copyright. For more information, see http://ojps.aip.org/rsio/rsicr.jsp
