Volcanic CO2 emissions play a key role in the geological carbon cycle and monitoring of volcanic CO2 fluxes helps to forecast eruptions. The quantification of CO2 fluxes is challenging due to rapid dilution of magmatic CO2 in CO2-rich ambient air and the diffuse nature of many emissions, leading to large uncertainties in the global magmatic CO2 flux inventory. Here, we report measurements using a new DIAL laser remote sensing system for volcanic CO2 (CO2DIAL). Two sites in the volcanic zone of Campi Flegrei (Italy) were scanned, yielding CO2 path-amount profiles used to compute fluxes. Our results reveal a relatively high CO2 flux from Campi Flegrei, consistent with an increasing trend. Unlike previous methods, the CO2DIAL is able to measure integrated CO2 path-amounts at distances up to 2000 m using virtually any solid surface as a reflector, whilst also being highly portable. This opens a new frontier in quantification of geological and anthropogenic CO2 fluxes.
Quantifying CO2 emissions from geological and anthropogenic sources is a key objective for Earth and environmental scientists. The geological carbon cycle1 buffers the CO2 content of the atmosphere and thereby global temperatures, through a balance between CO2 emissions from volcanoes and geological sources and CO2 take-up by temperature-dependent silicate rock weathering. This balance has maintained relatively stable atmospheric CO2 concentrations for the last 600 kyr2, thereby helping to stabilize temperatures on Earth3. Current estimates of global geological CO2 emissions have a large uncertainty4. The IPCC determine a geological CO2 flux of 370 Mt/yr5, while recent assessments of global CO2 flux indicate 540 Mt/yr4, neglecting potentially enormous but largely unquantified diffuse CO2 emission sources, such as that of the East African Rift, which alone may produce 71 Mt/yr6. Improving our quantitative understanding of geological CO2 emissions is therefore a fundamental goal for Earth scientists.
Geological CO2 emissions are the product of a complex exchange of CO2 between primordial mantle, mantle mixed with subducted material, continental crust and the exosphere. The degree to which CO2 emissions from arc volcanism reflects recycling of crustal CO2 or degassing of primordial mantle is an open question with large uncertainties7. Further measurements of the magnitudes of such fluxes can shed light on the CO2 recycling budget. Monitoring of gas emissions from volcanic areas is also useful for hazard assessment, as eruptive activity may be heralded by variations in gas composition and flux. CO2 is particularly useful as it is exsolved from magma at greater depths than other volatile species and therefore can reflect deep processes8,9,10,11.
It is clear, therefore, that volcanic CO2 flux measurement is of great importance in the Earth sciences, however, performing this is by no means straightforward. To quantify gas fluxes, concentration profiles are needed. Owing to the often moderate magmatic CO2 signal compared with the ambient CO2 concentration volcanic CO2 fluxes are usually derived indirectly. In-situ point samples of CO2/SO2 ratios using MultiGAS12, or remote sensing gas composition measurement with COSPEC13,14,15,16, or open path FTIR spectrometry (OP-FTIR)17,18 are combined with measurements of SO2 flux using ultraviolet remote sensing19,20. CO2 fluxes may also be determined using airborne in-situ CO2 sensors and flying vertical grids through volcanic plume cross-sections8,21,22. Diffuse soil degassing CO2 efflux is typically measured in situ as well, using accumulation chambers23,24,25, LICOR analysers26 or the eddy covariance method27,28. Degassing of magmatic CO2 often occurs over large areas (10 s to 100 s of m). This includes vented and diffuse releases, such as soil leakage and volcano flank degassing29,30. Such fluxes are typically highly variable in space and time31, which poses a challenge to sporadically applied measurement techniques. Estimating gas fluxes from in-situ point measurements may thus lead to large uncertainties32. Moreover, in situ acquisitions require significant time and effort, especially for time-lapse studies. Even if they can be performed rather swiftly, such as by airborne sampling, in situ measurements often bear risk, e.g. when flying or hiking near volcanic craters.
Remote sensing techniques are usually safer to deploy than in-situ techniques and can offer attractive spatial coverage, yielding 1-D (path averaged) or 2-D concentration profiles from which more robust flux estimates can be retrieved. OP-FTIR33,34,35, hyperspectral imagers36,37 and differential absorption spectroscopy (DOAS)38,39 have all been used to measure magmatic gas compositions and fluxes. Passive remote sensing approaches, however, depend on light coming from the end of the measurement path. This may limit their flexibility, e.g. to clear sky conditions. Moreover, one risks to sense photons that have not penetrated the volcanic plume40.
Active remote sensing techniques, including open path laser absorption spectrometers41,42, carry their own light source and overcome these difficulties. Ideally, such instruments would not require designed retroreflectors, as this severely limits the flexibility of the measurement setup, in particular for airborne acquisitions. Active remote sensing approaches without the need for retroreflectors, such as LIDAR, overcome all previously mentioned drawbacks. Instruments based on differential absorption LIDAR (DIAL) have been performing pioneering measurements of SO2 at volcanoes over 20 years ago43,44, yet have not become a standard approach to measure volcanic gases. DIAL is increasingly used in global warming science to measure path averaged, ambient CO2 concentrations45,46,47,48 and has even been used to map a CO2 plume at an industrial site49. Recently, a range-resolving DIAL has successfully measured 2-D volcanic CO2 concentration profiles in the Campi Flegrei volcanic zone (Italy) from which CO2 fluxes have been inferred50.
Perhaps the main reasons why LIDAR based active remote sensing has not established itself as a standard for volcanic gas measurement is that LIDAR systems are not readily portable, due to their size and power requirements46,47,49 and require extensive operational resources (personal, expertise, setup time, logistics, maintenance)44,47. A versatile volcanology tool, instead, requires high portability and flexibility, since many volcanoes are in remote and difficult to access locations. Quick setup times and efficient data acquisition are also highly beneficial to minimise exposure to hazards. In the last decade, technology has moved sufficiently far forward to allow for the development of systems that fulfil these requirements.
We have developed an active remote sensing instrument based on the differential absorption principle, called CO2DIAL, which is specifically designed to detect volcanic CO2 emissions. This article presents results of an experiment in which the CO2DIAL was used to sense volcanic CO2 at the Solfatara crater and the Pisciarelli fumaroles in the active volcanic area of Campi Flegrei (CF) in Italy (Fig. 1a). CF is a nested caldera, resulting from two large collapses that occurred ~39 ka BP and ~15 ka BP51,52. Being close to the city of Naples, CF poses a direct threat to approximately one million people53. During the measurements at CF the CO2DIAL demonstrated its key capabilities, quickly acquiring profiles of CO2 path-amounts both near (~100 m) and far (~1 km) from the source after only ~15 min setup, for both diffuse and vented emissions, whilst being highly portable and operated by a single person. The retrieved CO2 fluxes from CF confirm the unusually strong degassing activity previously observed26,54.
These results demonstrate how the CO2DIAL could provide effortless monitoring of magmatic CO2 fluxes for eruption monitoring purposes at a volcanic area at unrest such as CF. The CO2DIAL is rugged and has been transported on passenger airplanes as ordinary carry-on luggage, which makes rapid international deployment uncomplicated. The CO2DIAL is platform independent and can, for instance, be operated airborne, e.g., from a helicopter, airplane and eventually a UAV. As such the CO2DIAL fills a key operational gap in volcanic CO2 sensing techniques and opens a new, exciting route to knowledge about volcanic systems, in particular regarding magma dynamics and the geological carbon cycle.
The area of Pisciarelli encompasses a fault-related, fumarolic area, vigorously degassing water vapour and CO2, featuring a prominent, hot (90 to 110 °C) main vent and various smaller vents, including mud pools exhibiting gas bubbling. The Pisciarelli fumaroles therefore offer an optimal context for the CO2DIAL. The measurement geometry (Fig. 1b, see also Methods) mimics airborne acquisition and hence demonstrates the system’s capability for systematic airborne scanning of volcanic chains. The CO2DIAL was first placed in the vicinity of the Pisciarelli fumaroles (~60 m distance). During the measurement the optical axis of the instrument was pointing to the right of the centre of the plume (Fig. 1c) at a fixed angle. Strong, transient decreases in the grand ratio (GR, normalised intensity ratio, see Methods) are apparent (Fig. 2a), indicating the presence of considerable amounts of volcanic CO2 (Fig. 2b). The CO2 concentration at this angle was rapidly varying in a pulse like manner, which may be related to small scale wind eddies flushing in CO2 from a nearby vent. The path averaged CO2 concentrations (Equation (4), Methods) acquired with the CO2DIAL (between ~400 and 3163 ppm) agree well with those measured in situ during the acquisition of the data using a LICOR CO2 analyser (between 650 and 3000 ppm along the CO2DIAL path). Note that, in the following, CO2 concentrations measured by the CO2 DIAL always refer to path averaged concentrations. We found that working at a small distance between the instrument and the plume was impractical, since, due to the limited radial scanning velocity, scans took too much time. Depending on local, small-scale wind eddies in the plume area, opaque, condensed water vapour occasionally covered the measurement path. Sunlight scattered by these cloud then saturated the detector and made a repetition of the scan necessary. By moving further away we were able to perform a more rapid scanning of the plume area, reducing the chance of scattered sunlight interfering with the scan.
The CO2DIAL was located in a small parking lot located at an exposed part of a slope ~940 m northeast of Pisciarelli (Fig. 1b,d). The telescope was visually aligned with the Pisciarelli plume by means of a 90° flip mirror in the ocular tube of the telescope. The angular coverage of the scans was chosen generously at 7.5°, in order to include surroundings free of plume CO2 into the scan, to help unambiguously identify the plume. The optical power was set to 1 W (66% of maximum). Scans were performed by means of a step motor that pivoted the complete transmitter and telescope unit around the region of interest (Fig. 3a). The angular velocity of the scan was fixed at 0.525 mrad/s. A single scan took ~4 min out of which the plume itself was traversed in ~100 s. The CO2DIAL detected a strong increase in column integrated CO2 path length concentration product (hereafter called CO2 path-amount, Equation (3) in Methods). The latter varied from ~4 × 105 ppm.m up to 4.8 × 105 ppm.m near the centre of the plume, which, after dividing by the target range (measured by an onboard laser range-finder), corresponds to a path averaged concentration of up to 530 ppm (Fig. 3c). Assuming that this excess CO2 was contained entirely in the plume and estimating the plume diameter, permits us to estimate the in-plume concentration of CO2 as 3600 ± 530 ppm (1 SD, see Equation (5) in methods for derivation of the in-plume concentration), in good agreement with the LICOR in situ data (Fig. 3d).
Solfatara is a tuff cone associated with intense hydrothermal activity, manifesting fumaroles, hot soils and diffuse CO2 degassing. The Solfatara crater therefore offered an excellent possibility to probe a mixture between fumarole and diffuse degassing. An overview of the setup is given in Fig. 4a. The scanning angular velocity was 2.1 mrad/s. The two main fumarole vents, Bocca Nuova (BN) and Bocca Grande (BG), are located near the edge of the crater (Fig. 4b). In-situ measurements using the LICOR analyser that took place during the CO2DIAL scans confirmed high CO2 concentrations in their vicinity, as expected. Solfatara furthermore features zones of weaker degassing strength and an extensive diffuse soil CO2 output. The corresponding CO2 concentrations were clearly detectable with the LICOR analyser. A relatively high scanning inclination angle would have been necessary to scan over the main vents. This would, however, likely have missed excess CO2 concentrations in the centre of the crater (towards the instrument location) and further south along the rim (in situ points 1 to 4), in particular the weaker zones of anomalous CO2 release, both vented and diffuse. In addition, since we could visually observe that the plume water vapour movement was a superposition of vertical flow and a horizontal component towards south/southeast, it was clear that the plume CO2 would be advected away from BN and BG towards the southeast (roughly along the crater rim). We therefore chose a scanning inclination such that the scan altitude at BN was 2 m above the ground, which while missing the direct observation of the BG vent, would capture the CO2 emissions from the BG vent downwind. From Fig. 5 it can be seen that the CO2DIAL detected high CO2 concentrations near 35° that are compatible with the LICOR in situ values, both having a symmetric profile. However, with respect to the in situ values these CO2 concentrations are shifted by 9°, i.e. ca. 20 m further south. This can be explained by the fact that the in situ measurements at points 5 to 8 were performed at slightly higher elevation (Fig. 4b) and were not covered by the CO2DIAL measurement path during the near field scans. At 60° another zone of higher CO2 concentrations can be identified in all three scans (Fig. 5a–c), which reveals the presence of a diffuse CO2 source. Given the associated position, the increase in CO2 concentration near 50° (Fig. 5a–c) could be related to BC, a vent that appeared recently26. Lastly, there are numerous, persistent small-scale fluctuations in the profiles (Fig. 5a–c), such as near 16°, confirming widespread CO2 degassing activity.
To further test the results from the CO2LIDAR, we scanned the same degassing feature from a location further away, with the instrument installed ~550 m northwest of the main vents (Fig. 4a). The inclination for these measurements included both BN and BG (Fig. 4b). The detected CO2 concentration versus the heading angle appears smoother (Fig. 6). This could be partly due to the slightly higher inclination (beam travelled ~5 m above ground near BN) and larger atmospheric contribution which both tend to diminishes the relative contribution of CO2 from weaker vents. Moreover, using the same angular scan velocity of 2.1 mrad/s as before the probed sector size was ~1.2 m/s compared with 30 cm/s for the near field scan. As a result, the effect of the CO2 is averaged, contributing to a smoother profile. Out of all scans performed from the Solfatara far field position, the maximum in-plume volcanic CO2 concentration detected with the CO2DIAL was 1900 ± 303 ppm.
Our CO2 path-amount profiles together with estimated vertical plume transport speeds were used to retrieve CO2 fluxes as described in the Methods section. Table 1 summarizes the results for the scans carried out at Pisciarelli and Solfatara during the complete duration of the experiment. For Solfatara the sectors considered for flux retrieval include the area around the main vents as shown in Fig. 4b. As already suggested by the CO2 concentrations (Fig. 5), the CO2 fluxes vary strongly during the course of even tens of minutes (Table 1). The time-averaged CO2 flux estimated for Pisciarelli is 3.1 ± 2.5 kg/s (1 SD) or 266 ± 212 tons/day. The average over subsequent scans yields a mean CO2 flux for Solfatara of 9.3 ± 4.7 kg/s (805 ± 408 tons/day) for the near field and 7.7 ± 4.5 kg/s (664 ± 386 tons/day) for the far field scans, associated with a total average of 715 ± 394 tons/day for Solfatara.
We have used the newly developed CO2DIAL to sense volcanic CO2 at the restless Campi Flegrei volcanic zone in Italy. The CO2 concentrations measured with the CO2DIAL are in good agreement with in situ values acquired with a LICOR CO2 analyser. Care has to be taken when comparing in situ values with those from the CO2DIAL, since the in situ measurements detect contributions from a point-like volume around the instrument, whereas concentrations detected by the CO2DIAL correspond to contributions from points all along the column between instrument and hard target (Methods). The most robust comparison is produced when the CO2 concentration in the column is governed by the same CO2 source sensed by the in situ device. For Solfatara, this was the more the case for near field scans (Fig. 5) than for the far field scans (Fig. 6).
CO2 path-amount profiles have been converted into fluxes using the vertical plume transport speed estimated from video footage (Methods). The precision of the CO2 concentration along with error of the vertical CO2 transport speed propagate to an error of the CO2 flux (Equation (12), Methods), which is on average 60% (mean over relative errors of all computed fluxes). The system has yet not reached its full potential, mainly in terms of precision. Future work will focus on increasing precision to be able to sense more subtle volcanic CO2 signals and increase confidence of the retrieved CO2 fluxes. Further, the plume speed measurement technique will be improved to reduce the related uncertainty.
While much variability in the measured CO2 path amounts is the product of local fluctuations in transport speed and direction within the plume, the varying CO2 path-amount detected at Pisciarelli (Fig. 2) may reflect a varying fumarole degassing rate on a timescale of tens of seconds42. A further, more systematic investigation, including a frequency analysis, would enable improved identification of the origin of the variability. This represents one of the doors opened by new active remote sensing tools with high temporal resolution, such as the CO2DIAL. A longer time-scale variability in CO2 flux is observed at both Pisciarelli and Solfatara. Table 1 suggests that strong variations of the CO2 flux by 20 to 76% may occur within less than an hour, which is consistent with previous observations50 which found average degassing rates at Pisciarelli of 227 tons/day (mean of all measurements) in late 2014. That result is slightly lower than the average of 266 ± 212 tons/day found in this study. Results from more previous years are significantly lower (e.g., ~180 tons/day in 201242). The high quality dataset of CO2 fluxes captured by the Osservatorio Veusiviano geochemistry team has revealed that fluxes have risen by ~40% every year between 2008 and 201455. The present finding suggests an increase in CO2 flux by a factor of 266/227 (17%) in 1.5 years, which is reasonably consistent with previous observations and certainly indicates a general trend of increasing CO2 efflux from Pisciarelli. This contributes to the general picture of increased degassing associated with enhanced ground uplift observed in the area26,54.
For the mixed degassing (vented and diffuse) at Solfatara it is more challenging to compare with in-situ fluxes as the boundaries of the area investigated are generally not exactly the same. This is already the case for the near and far field Solfatara scans of the present study, which span different areas. Yet, the CO2 fluxes for the far and the near field agree within error. Note that the CO2DIAL measures path-amounts, (Methods) with contributions from vented and diffuse CO2 emissions all across the measurement line. Therefore, differences between the fluxes from the far and near field measurement are very likely due to the different integrated soil areas covered by the near and the far field scans. The average flux of 715 tons/day found in this study is considerably larger than CO2 fluxes from previous measurements between 2012 and 2013 at the main vents (BN and BG), which yielded between 251 and 300 tons/day26,42. However, these values account for the fumarole degassing, excluding diffuse degassing. The fumarole emissions near BG and BN were estimated to correspond to roughly 40% of the total soil CO2 output26, resulting in a total flux of ~1560 tons/day, however, for the whole Solfatara/ Pisciarelli area. In line with this value, a diffuse degassing rate of ~1500 tons/day for a 0.5 km2 area of Solfatara has been reported previously56. It is hard to scale the calculated flux in this study up to the size of the whole degassing area of CF as there are areas with a very low flux and areas with a very high flux. We can, however, provide an estimation for the Solfatara crater alone, excluding the area of the Pisciarelli fumaroles. For the Solfatara far field scans, the sector accounted for in the flux computation of this work corresponds to a surface of ~20000 m2 inside the Solfatara crater (region of interest, shaded area in Fig. 4a,b). When integrating over all headings (dashed triangle, Fig. 4a), as opposed to the region of interest only, the flux doubles on average, amounting to ~1400 tons/day. This larger region has an area of ~43000 m2, which corresponds to ~14% of the crater area of ~300000 m2, approximately bounded by the dashed white line in Fig. 4a. Flux maps of previous in situ studies57,58 suggest that the area covered in our study accounts for 31% (=43000/140000) to 61% (=43000/70000) of the anomalous degassing area inside the crater. Assuming a consistent flux per unit area throughout the crater implies a CO2 flux between 2300 and 4600 t/day for the Solfatara crater alone, including the contribution of the main fumaroles (BN and BG) and other minor vented emissions on the southeast wall of the crater. In the future, a more spatially comprehensive flux measurement is envisaged, scanning the whole of Solfatara cater, which will allow to compare fluxes of the Solfatara area with greater confidence.
Whilst smaller than at Pisciarelli, Solfatara has seen a steady increase in CO2 degassing rate in the last decade or so55. Recent findings59,60 indicate that apart from the vented degassing area near BN and BG there is a “hot spot” zone of approximately rectangular shape along the south-eastern edge of the Solfatara crater (between 0 and 36° in Fig. 5), featuring high anomalous CO2 release, in line with the LICOR data of this study. It underpins the high CO2 concentrations detected in that zone by the CO2DIAL and furthermore supports that the associated CO2 amounts largely correspond to diffuse emissions and minor vents with formerly unknown CO2 output.
In summary, the CO2 fluxes found with the CO2DIAL are compatible with previous results, indicating strong, increasing CO2 degassing at CF. To the best of our knowledge, it is the first time that a man-portable active remote sensing system has been used to acquire volcanic CO2 profiles. The outcomes show the tremendous potential of a portable gas remote sensing system. The ease and flexibility with which the presented data has been attained represents a breakthrough in the measurement of volcanic CO2. The results show that the CO2DIAL is a powerful tool to overcome the aforementioned drawbacks of in situ measurement techniques by offering a faster, safer and comprehensive acquisition (spatial coverage yields CO2 concentration profiles). The platform demonstrated its capability to execute effortless time-lapse acquisitions within minutes or tens of minutes, depending on the area covered and the target range. This is a very desirable feature to analyse correlation between CO2 degassing and uplift at inhabited volcanic zones, such as CF, where days can be a crucial time scale.
In conclusion, a man-portable active remote sensing tool as the CO2DIAL represents a step change in volcanology by opening completely new measurement possibilities. One of these prospects is uncomplicated, systematic probing of arc volcanoes to populate the global volcanic CO2 flux data inventory and help constraining the geological carbon cycle and deliver new insights in the magmatic plumbing system of subaerial volcanoes. Applications of the CO2DIAL also include more general areas of environmental and Earth system science.
The CO2DIAL is based on continuous wave, hard target differential absorption LIDAR (DIAL), measuring path averaged CO2 concentrations. A DIAL transmits and receives photons at two wavelengths, one corresponding to a rotational-vibrational transition of CO2 where photons are being absorbed (λON) and one where there is no associated absorption (λOFF). λON and λOFF are closely spaced to each other so that atmospheric propagation effects other than absorption by the probed molecule can assumed to be negligible. By taking the ratio of the optical powers associated with the received signals for λON and λOFF one arrives at47
where R is distance between the instrument and the hard target (e.g. the soil), Δσ is the difference between the molecular absorption cross sections of CO2 associated with λON and λOFF, P(λ) is the received power (hereafter referred to as science signal as it carries the information of scientific interest) and P(λ)ref is the transmitted optical power (referred to as reference signal). The latter is measured as a reference to normalize fluctuations of the transmitted power. Δτ is the differential optical depth. The ratio of the normalized powers in Equation (1) is referred to as the grand ratio (GR). The presence of excess CO2 leads to a decrease in GR (Fig. 2a). By measuring the GR as well as R one may obtain the path averaged CO2 number density , which for a horizontal measurement path (as in this work) can be converted to a concentration (in ppm) as
where Nair is the (moist) air number density computed with knowledge of specific humidity, air pressure and temperature.
Supplementary Fig. S1 presents an overview of the CO2DIAL instrument as used for this work, which is an all-fiber optic instrument. The instrument is a prototype with development costs of the order of 100 k €. A next generation commercial prototype is planned with associated costs roughly a third of that. The CO2DIAL is designed to measure path averaged CO2 concentrations for path lengths up to 2000 m. The instrument has been developed to acquire volcanic CO2 concentration profiles from which CO2 fluxes can be deduced. The complete system weighs ~26 kg and uses a power of ~20 W in idle mode and ~70 W when the laser is emitting at its full optical power. The CO2DIAL can operate for ~1 h on a single 600 g, 5000 mAh lithium polymer battery. The laser consists of two continuous wave distributed feedback (DFB) fiber seed lasers emitting at λON = 1572.992 nm and λOFF = 1573.173 nm61. The seed laser beams are amplitude modulated using two LiNbO3 electro-optical modulators (EOM) at a respective sine tone near 5 kHz and simultaneously amplified by an Erbium doped fiber amplifier (EDFA) before being transmitted. Light backscattered by a hard target is received by a commercial 200 mm diameter Schmidt-Cassegrain Telescope with a focal length of 1950 mm. The analogue to digital converter (ADC) operates at 250 kSamples s−1 and has a resolution of 16-bit. The integration time per scan angle was set to 4000 EOM modulation periods, which corresponds to chunks of length 784 ms for both science and reference channel. Each of these chunks of data is demodulated using a digital lock-in routine (or through performing a FFT for diagnostic purposes) written in National Instruments LabVIEW. This is done in real-time for quality control purposes (e.g. to check for abnormal baseline drift) and again in the post processing of the data to compute CO2 concentrations. After the lock-in operation one arrives at four DC signals associated with P(λON), P(λOFF), P(λON)ref and P(λOFF)ref. The GR and Δτ are calculated using the right hand side of Equation (1) after taking the mean of each of the four signals. Each point in Figs 2, 3, 5 and 6 is associated with an integration time of 784 ms. The target distance R is measured by a range finder (DLEM, Jenoptik, Germany), based on a 1550 nm LIDAR with pulse energy of 500 μJ and accuracy <1 m. The working principle of the range finder is the time of flight measurement. The integration time was set to 500 ms.
CO2 concentration retrieval
DIAL instruments generally have an offset, i.e. Δτ(R = 0) ≠ 0. In addition, due to the fact that the metastable state lifetime of the erbium atoms is comparable to the period of the seed laser amplitude modulation, the EDFA distorts the amplitudes of the seed laser signals. Demodulation of the distorted science and reference signals creates an offset that depends on the signal power, but which is repeatable. As a result, the measured GR are shifted with respect to those expected from Equation (1). Therefore, to retrieve path averaged CO2 concentrations a calibration is performed using the ambient CO2 at the measurement site, away from any CO2 degassing (Supplementary Fig. S2). To this end, GR were measured in the well-mixed ambient atmosphere for various ranges, converted to differential optical depths using Equation (1) and plotted against the CO2 path length concentration product (path-amount, product of ambient CO2 concentrations and ranges). A straight line,
was fitted to the ensemble of points using a least square algorithm. Equation (3) was used as the calibration curve, with slope a (in ppm.m−1) and instrumental off set b. Employing the calibration curve the measured Δτat CF were converted to CO2 path-amounts (in ppm.m), which were used for further computations as detailed below.
Conversion of CO2 path-amounts to in-plume CO2 concentrations
Figure 7 shows a simplified sketch of a typical measurement geometry at Campi Flegrei. Since the CO2DIAL yields CO2 path-amounts (ppm.m), volume CO2 concentrations measured in situ (in the plume area) cannot be compared directly with the result from the CO2DIAL. If the plume extension approximately equals the CO2DIAL path length, dividing column amounts by the path length may be done, i.e.
such as shown in Figs 2b and 5, yielding path averaged CO2 concentrations. Of course, for path lengths much longer than the plume extension cannot be compared with in situ values taken in the plume. In these cases CO2 column amounts were converted to in-plume concentrations , which account for the volcanic and the ambient CO2 (such as detected by the LICOR analyser). They were obtained using
where is the volcanic excess CO2 path-amount (in ppm.m), obtained as
where = 380 ppm (ambient CO2 concentration at CF from in situ measurement) and R is the target range.dpl is the lateral dimension of the plume, calculated as
where Δri is the lateral distance increment (in m) per angular step i. with the range per angle Ri (see Fig. 7), the angular scan velocity and the time interval Δti between points (angles), retrieved from the time stamps of the data chunks.
Identifying the plume
An increase in volcanic CO2 would lead to an increased column averaged amount of CO2 and would thus decrease the GR (increase Δτ). When performing a scan across a sufficiently strong CO2 emission this behaviour occurs as a function of scanning angle. By repeatedly scanning across the emission feature, artefacts can be ruled out and the decrease in GR as well as the angle where this happens can be unambiguously attributed to CO2 originating from anomalous degassing activity. Of course, this assumes that the plume extension is sufficiently stable between subsequent scans. For vented degassing, the position of the telescope (and hence the angle) when it entered an anomalously degassing zone could be determined visually using the telescope. This procedure works well when the scanned horizon also contains areas with ambient atmosphere only. This was not the case for the near field Solfatara scans since the instrument was placed inside the anomalous degassing (i.e. the main plume) itself already. Nonetheless, one could identify a rather stable symmetric increase in CO2 concentration, which was chosen as the region of interest to integrate over to arrive at the CO2 flux.
To compute the mass flux of CO2 (in kg/s) the volcanic excess CO2 path-amounts (Equation 6) were integrated over the lateral plume extension as
where vpl is the plume transport speed (m/s), more precisely the component of the transport speed vector perpendicular to the scanned cross section. It was assumed constant during the scan and across the plume cross sectional area (see section plume speed estimation). Since the scans were performed horizontally, here it refers to the magnitude of the vertical component of the transport speed vector (Fig. 7). Nair is the number density of moist air, computed using meteorological data (pressure, temperature, humidity) acquired by a portable meteo station close to the instrument (Fig. 7). is the molar mass of CO2 (kg mol−1) and NA is Avogadro’s constant (mol−1). For a vented emission, as at Pisciarelli, the integration boundaries were determined by the angles between which there was a steady increase followed by steady decrease in CO2 concentration above the background. For Solfatara, the integration was carried out for the region of interest around the main vents, that is, between in situ measurement points 8 and 4 (near field) and 8 and 2 (far field) (Fig. 4a,b and section identifying the plume). This interval was chosen since the scans indicated that it contained the CO2 from the main fumaroles as well as CO2 from along the crater edge between in situ point 5 and 8, discernible by a repeatable feature (Fig. 5).
Plume speed estimation
vpl may be estimated from the wind vector. However, particularly close to the ground or close to reliefs, such as at Solfatara, the topography gives rise to a complicated wind vector field. The wind speed may thus be quite different from the actual transport speed of the plume. Further, the meteo station was not measuring the wind field components, but the true wind direction. For the near field Solfatara measurements the wind veered around 285° (mean of wind speeds for all scans). For the Solfatara far field measurement on 02/03/2016 there was no wind direction data available as the compass was not calibrated. For the Solfatara far field measurements on 03/02/2016 the mean wind direction was 168°. For the scans at Pisciarelli the wind direction measured was on average 200°, which means the Pisciarelli fumaroles were in the wind shadow of the Solfatara crater rim. vpl was estimated therefore directly from the plume movement. It was assumed that the movement of the visible condensed plume water vapour aerosol propagated with the same velocity as the volcanic CO2. Video footage with a resolution of 480 × 640 pixels2 was acquired from the plume using a commercial digital camera placed next to the CO2DIAL telescope. During each set of scans a video clip of the plume was produced (4 in total). A video analysis program (Tracker from Open Source Physics) was used to analyse the video footage. At a given video frame a pixel of the plume was picked (Supplementary Fig. S3a). The pixel (which corresponds to a parcel of gas) was moving with each frame until it eventually experienced a decrease in brightness, usually after 3 to 13 seconds, due to associated water droplets evaporating. The analysis focused on pixels within and near the telescope field of view. Transformation from pixel into real world coordinates was carried out via distance calibration using the size of a real world feature (e.g. a house). These features were chosen such that they were aligned approximately parallel to the plane of the plume. In cases where these features were not perpendicular to the telescope line of sight (for Pisciarelli) their projected size was used, employing Google Earth software. The camera line of sight was usually quasi perpendicular to the (on average vertical) plume transport direction (for Pisciarelli, cos θ = 0.997 in Fig. 7). For the Solfatara scans (both near and far field) the plume speed had a horizontal component, which was, however, quasi perpendicular to the telescope viewing direction. This resulted therefore in a negligible length projection error. Frames were converted to time using the frame rate of the camera (30 Hz). The tracking software enabled to extract the relative vertical position of a pixel (y-position). vpl was derived from the slope of the regression line fitted on the y-positions (Supplementary Fig. S3b). The error was extracted from the root mean square error of the fit (RMSE of observed minus modelled y-coordinates). For a tracked parcel of gas this error accounts for fluctuations in transport speed due to a complex, perhaps turbulent wind field. Note that vpl derived in that fashion represents an average speed over the course of the measurement. The RMSE represents an uncertainty over time, but for only a part of the plume. The procedure was therefore repeated for various parcels across the plume. The retrieved speeds and their RMSEs are summarized in Supplementary Table S1. Comparing the derived speeds provided a way to assess fluctuations in vertical transport speed across the plume as viewed by the CO2DIAL telescope. Assuming normally distributed plume speeds the associated standard deviation was computed using Student’s t distribution62, developed for small number of samples, which are, however, statistically meaningful. Here, 8 tracks have been performed per video (Supplementary Table S1). Note that each of the 8 plume speeds per video is based on a fit of >100 points and thus exhibit a certain robustness. The uncertaintyΔvpl of vpl was estimated as the maximum out of the Student’s SD and the mean of the RMSEs (Supplementary Table S1). For the scans carried out at far field Solfatara at the evening of 02/03/2016, y-positions versus time were usually quite steady, suggesting a rather laminar plume flow, which resulted in a lower RMSE (Supplementary Fig. S3b and Table S1). Supplementary Fig. S3c,d show two less steady y-tracks of parcels of the far field Solfatara scans the morning after, suggesting a higher degree of turbulence and hence a higher associated uncertainty of the plume transport speed.
Estimating measurement precision
The total relative uncertainty of the CO2 path-amount was evaluated as
with the signal-to-noise-ratio (SNR)
where 〈GR〉 and σGR are the mean and standard deviation of the grand ratio values, respectively. They were estimated from time series acquired at fixed angles in between the scans at CF. The SNR accounts for all noise sources occurring during acquisition, including instrumental noise, non-stationary baseline drift, solar background noise, atmospheric noise (mostly air turbulence) and perturbation by aerosol scattering. It is therefore an indicator of the instrument’s true measurement performance in the field63. The second term quantifies the relative range error (standard deviation of ranges σR over mean of ranges 〈R〉), which is typically ~1 m. The relative uncertainty due to hard target speckle was estimated as64
where D is the spot diameter on the hard target (in m) and η is the dimension of the telescope FOV (in m) on the hard target.
The total relative error of the CO2 flux was computed as
How to cite this article: Queiβer, M. et al. A new frontier in CO2 flux measurements using a highly portable DIAL laser system. Sci. Rep. 6, 33834; doi: 10.1038/srep33834 (2016).
Berner, R. A. A model for atmospheric CO2 over Phanerozoic time. Am. J. Sci. 291, 339–376, 10.2475/ajs.291.4.339 (1991).
Zeebe, R. E. & Caldeira, K. Close mass balance of long-term carbon fluxes from ice-core CO2 and ocean chemistry records. Nature Geosci. 1, 312–315 (2008).
Liu, Z., Dreybrodt, W. & Liu, H. Atmospheric CO2 sink: Silicate weathering or carbonate weathering? Appl. Geochem. 26, 292–294 (2011).
Burton, M., G. Sawyer, G. & Granieri, D. Deep carbon emissions from volcanoes. Rev. Mineral. Geochem. 75, 323–354 (2013).
IPCC. Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change [ Stocker, T. F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S. K., Boschung, J., Nauels, A., Xia, Y., Bex, V. & Midgley, P. M. (eds)]. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA, 1535 pp, 10.1017/CBO9781107415324 (2013).
Lee, H. et al. Massive and prolonged deep carbon emissions associated with continental rifting. Nature Geosci. 9, 145–149 (2016).
Kelemen, P. B. & Manning, C. E. Reevaluating carbon fluxes in subduction zones, what goes down, mostly comes up. PNAS 112, 3998–4006, 10.1073/pnas.1507889112 (2015).
Gerlach, T. M. et al. Application of the LI-COR CO2 analyzer to volcanic plumes: A case study, volcán Popocatépetl, Mexico, June 7 and 10, 1995. J. Geophys. Res. 102, 8005–8019 (1997).
Carapezza, M. L., Inguaggiato, S., Brusca, L. & Longo, M. Geochemical precursors of the activity of an open-conduit volcano: The Stromboli 2002–2003 eruptive events. Geophys. Res. Lett. 31, L07620 (2004).
Aiuppa, G. et al. First observational evidence for the CO2-driven origin of Stromboli’s major explosions. Solid Earth 2, 135–142 (2011).
Poland, M. P., Miklius, A., Sutton, J. & Thornber, C. R. A mantle-driven surge in magma supply to Kīlauea Volcano during 2003–2007, Nature Geosci. 5, 295–300 (2012).
Shinohara, H. A new technique to estimate volcanic gas composition: plume measurements with a portable multi-sensor system. J. Volcanol. Geotherm. Res. 143, 319–333 (2005).
Allard, P. et al. Eruptive and diffuse emissions of CO2 from Mount Etna, Nature 351, 387–391 (1991).
Symonds, R. B., Reed, M. H. & Rose, W. I. Origin, speciation and fluxes of trace-element gases at Augustine volcano, Alaska – Insights into magma degassing and fumarolic processes. Geochim. Cosmochim. Acta 56, 633–657, 10.1016/0016-7037(92)90087-Y (1992).
Gerlach, T. M., McGee, K. A., Sutton, A. J. & Elias, T. Rates of volcanic CO2 degassing from airborne determinations of SO2 Emission rates and plume CO2/SO2: test study at Pu′u ′O′o Cone, Kilauea Volcano, Hawaii. Geophys. Res. Lett. 25, 2675- 2678 (1998).
Shinohara, H., Kazahaya, K., Saito, G., Fukui, K. & Odai, M. Variation of CO2/SO2 ratio in volcanic plumes of Miyakejima: Stable degassing deduced from heliborne measurements. Geophys. Res. Lett. 30, 1208, 10.1029/2002GL016105 (2003).
Allard, P., Burton, M. & Muré, P. Spectroscopic evidence for a lava fountain driven by previously accumulated magmatic gas, Nature 433, 407–410 (2005).
Burton, M., Allard, P., Muré, F. & La Spina, A. Magmatic gas composition reveals the source depth of slug-driven Strombolian explosive activity, Science 317, 227–230 (2007).
Galle B., Oppenheimer C., Geyer A., McGonigle A. & Edmonds M. A miniaturised ultraviolet spectrometer for remote sensing of SO2 fluxes: a new tool for volcano surveillance. J. Volcanol. Geotherm. Res. 119, 241–254 (2002).
Fischer, T. et al. The 2005 and 2006 eruptions of Ol Doinyo Lengai: assessing deep and shallow processes at an active carbonatite volcano using volatile chemistry and fluxes. Eos Trans. AGU 87, Fall Meet. Suppl., Abstract V14B-04 (2006).
Wardell, L. J., Kyle, P. R. & Chaffin, C. Carbon dioxide and carbon monoxide emission rates from an alkaline intra-plate volcano: Mt. Erebus, Antarctica. J. Volcanol. Geotherm. Res. 131, 109–121 (2004).
Diaz, J. A. et al. Utilization of in situ airborne MS-based instrumentation for the study of gaseous emissions at active volcanoes. Int. J. Mass. Spectrom. 295, 105–112 (2010).
Lewicki, J. et al. Comparative soil CO2 flux measurement and geostatistical estimation methods on Masaya volcano, Nicaragua. Bull. Volc. 68, 76–90 (2005).
Granieri, D. et al. Atmospheric dispersion of natural carbon dioxide emissions on Vulcano Island, Italy. J. Geophys. Res. Solid Earth 119, 5398–5413 (2014).
d’Alessandro, W. D. et al. Diffuse and focused carbon dioxide and methane emissions from the Sousaki geothermal system, Greece. Geophys. Res. Lett. 33, L05307 (2006).
Aiuppa, A. et al. First observations of the fumarolic gas output from a restless caldera: Implications for the current period of unrest (2005–2013) at Campi Flegrei. Geochem. Geophys. Geosyst. 14, 4153–4169 (2013).
Werner, C., Wyngaard, J. C. & Brantley, S. L. Eddy-Correlation Measurement of Hydrothermal Gases, Geophys. Res. Lett. 27, 2925–2928 (2000).
Werner, C., Chiodini, G., Voigt, D. E. & Brantley, S. L. Monitoring volcanic hazard using eddy covariance at Solfatara volcano, Naples, Italy, Earth Planet. Sci. Lett. 210, 561–577 (2003).
Chiodini, G., Cioni, R., Giudi, M., Raco, B. & Marini, L. Soil CO2 flux measurements in volcanic and geothermal areas. Appl. Geochem. 13, 543–552 (1998).
Hards, V. L. Volcanic contributions to the global carbon cycle. British Geological Survey Occasional Report No. 10, 26pp. (2005).
Cardellini, C. et al. Accumulation chamber measurements of methane fluxes: application to volcanic-geothermal areas and landfills. Appl. Geochem. 18, 45–54 (2003).
Aiuppa, A. et al. Rates of carbon dioxide plume degassing from Mount Etna volcano. J. Geophys. Res. 111, B09207 (2006).
Naughton J. J., Derby, J. V. & Glover, R. B. Infrared measurements on volcanic gas and fume: Kilauea eruption, 1968. J. Geophys. Res. 74, 3273–3277 (1969).
Burton, M. R., Oppenheimer, C., Horrocks, L. A. & Francis, P. W. Remote sensing of CO2 and H2O emission rates from Masaya volcano, Nicaragua. Geology 28, 915–918 (2000).
La Spina, A. et al. New insights into volcanic processes at Stromboli from Cerberus, a remote-controlled open-path FTIR scanner. J. Volcanol. Geotherm. Res. 249, 66–76 (2013).
Chen, H. S. Space Remote Sensing Systems – An Introduction. Academic Press Inc., Orlando, Florida, USA (1985).
Spinetti C., Carrere, V., Buongiorno, M. F., Sutton, A. J. & Elias, T. Carbon dioxide of Pu’u’O’ o volcanic plume at Kilauea retrieved by AVIRIS hyperspectral data. Remote Sens. Environ. 112, 3192–3199 (2008).
Platt, U. Differental optical absorption spectroscopy (DOAS). Chem. Anal. Series 127, 27–83 (1994).
Weibring, P. et al. Monitoring of volcanic sulphur dioxide emissions using differential absorption lidar (DIAL), differential optical absorption spectroscopy (DOAS) and correlation spectroscopy (COSPEC). Appl. Phys. B 67, 419–426 (1998).
Kern, C. et al. Improving the accuracy of SO2 column densities and emission rates obtained from upward-looking UV-spectroscopic measurements of volcanic plumes by taking realistic radiative transfer into account. J. Geophys. Res. 117, D20302, 10.1029/2012JD017936 (2012).
Menzies, R. T. & Chahine, M. T. Remote Atmospheric Sensing with an Airborne Laser Absorption Spectrometer. Appl. Opt. 13, 2840–2849 (1974).
Pedone, M. et al. Volcanic CO2 flux measurement at Campi Flegrei by Tunable Diode Laser absorption Spectroscopy. Bull. Volc. 76, 10.1007/s00445-014-0812-z (2014).
Carapezza, M. et al. Measurement of volcanic SO2 in the atmosphere by DIAL remote sensing techniques. in Man and His Ecosystem, edited by Brasser, L. J. & Mulder, W. C. pp. 77–82, Elsevier, New York (1989).
Edner, H. et al. Total Fluxes of Sulfur Dioxide from the Italian Volcanoes Etna, Stromboli and Vulcano Measured by Differential Absorption LIDAR and Passive Differential Optical Absorption Spectroscopy. J. Geophys. Res. 99, 827–838 (1994).
Gibert, F., Flamant, P. H., Bruneau, D. & Loth, C. Two- micrometer heterodyne differential absorption lidar measurements of the atmospheric CO2 concentration in the boundary layer. Appl. Opt. 47, 944–956 (2008).
Amediek, A., Fix, A., Wirth, M. & Ehret, G. Development of an OPO system at 1.57 μm for integrated path DIAL measurement of atmospheric carbon dioxide. Appl. Phys. B 92, 295–302 (2008).
Abshire, J. B. et al. Pulsed airborne lidar measurements of atmospheric CO2 column absorption. Tellus B 62B, 770–783 (2010).
Sakaizawa, D. et al. An airborne amplitude-modulated 1.57 μm differential laser absorption spectrometer: simultaneous measurement of partial column-averaged dry air concentration of CO2 and target range. Atmos. Meas. Tech. 6, 387–396 (2013).
Robinson, R. A. et al. First measurements of a carbon dioxide plume from an industrial source using a ground based mobile differential absorption lidar. Environ. Sci.: Processes Impacts 16, 1957–1966 (2014).
Aiuppa, A. et al. New ground-based lidar enables volcanic CO2 flux measurements. Sci. Rep. 5, 13614; 10.1038/srep13614 (2015).
Orsi, G., D’ Antonio, M., de Vita, S. & Gallo, G. The Neapolitan yellow tuff, a large-magnitude trachytic phreatoplinian eruption; eruptive dynamics, magma withdrawal and caldera collapse. J. Volcanol. Geotherm. Res 53, 275- 287 (1992).
Costa, A. et al. Quantifying volcanic ash dispersal and impact of the Campanian Ignimbrite super-eruption. Geophys. Res. Lett. 39, L10310 (2012).
Ricci, T., Barberi, F., Davis, M. S., Isaia, R. & Nave, R. Volcanic risk perception in the Campi Flegrei area. J. Volcanol. Geotherm. Res. 254, 118- 130 (2013).
Chiodini, G. et al. CO2 degassing and energy release at Solfatara volcano, Campi Flegrei, Italy. J. Geophys. Res. 106, 213–221 (2001).
d’Auria, L. Update sullo stato dei Campi Flegrei, INGV, Sezione di Napoli, Report, available at: ftp://ftp.ingv.it/pro/web_ingv/Convegno_Struttura%20_Vulcani/presentazioni/15_D’auria_CampiFlegrei/dauria_cf.pdf, accessed in March 2016 (2015).
Granieri, D., Chiodini, G., Marzocchi, W. & Avino, R. Continuous monitoring of CO2 soil diffuse degassing at Phlegraean Fields (Italy): influence of environmental and volcanic parameters. Earth Planet. Sci. Lett. 212, 167–179 (2003).
Chiodini, G. et al. Long-term variations of the Campi Flegrei, Italy, volcanic system as revealed by the monitoring of hydrothermal activity. J. Geophys. Res. 115, B03205, 10.1029/2008JB006258 (2010).
Granieri, D., Costa, A., Macedonio, G., Bisson, M. & Chiodini, G. Carbon dioxide in the urban area of Naples: Contribution and effects of the volcanic source. J. Volcanol. Geotherm. Res. 260, 52–61 (2013).
Tassi, F. et al. Diffuse soil emission of hydrothermal gases (CO2, CH4 and C6H6) at the Solfatara crater (Phlegraean Fields, southern Italy). Appl. Geochem. 35, 142–153 (2013).
Bagnato, E. et al. First combined flux chamber survey of mercury and CO2 emissions from soil diffuse degassing at Solfatara of Pozzuoli crater, Campi Flegrei (Italy): Mapping and quantification of gas release. J. Volcanol. Geotherm. Res. 289, 26- 40 (2014).
Rothman, L. S. et al. The HITRAN 2012 molecular spectroscopic database. J. Quant. Spectrosc. Radiat. Transfer 130, 4–50 (2013).
Student, The probability error of a mean. Biometrika 6, 1–25 (1908).
Queißer, M., Burton, M. & Fiorani, L. Differential absorption lidar for volcanic CO2 sensing tested in an unstable atmosphere. Opt. Express 23, 6634–6644 (2015).
MacKerrow, E. P., Schmitt, M. J. & Thompson, D. C. Effect of speckle on lidar pulse–pair ratio statistics. Appl. Optics 36, 8650–8669 (1997).
Tarquini, S. et al. TINITALY/01: a new Triangular Irregular Network of Italy. Ann. Geophys. 50, 407–425 http://tinitaly.pi.ingv.it (2007).
The research leading to these results has received funding from the European Research Council under the European Union’s Seventh Framework Programme (FP/2007-2013)/ERC Grant Agreement n. 279802. Our gratitude goes to Rosario Avino and Antonio Carandente (INGV Osservatorio Vesuviano), who sampled the in situ CO2 concentrations. We thank Graham Allen (NASA Goddard) and Luca Fiorani (ENEA) for sharing extremely valuable experiences in LIDAR development with us.
The authors declare no competing financial interests.
Electronic supplementary material
About this article
Cite this article
Queiβer, M., Granieri, D. & Burton, M. A new frontier in CO2 flux measurements using a highly portable DIAL laser system. Sci Rep 6, 33834 (2016). https://doi.org/10.1038/srep33834
This article is cited by
Scientific Reports (2019)
Scientific Reports (2019)
Scientific Reports (2017)