Impact of the aerosol type on HICO TM atmospheric correction in coastal waters

Introduction Conclusions References


Introduction
The radiative impact of aerosol is a key issue in the atmospheric radiative transfer models devoted to the earth/atmosphere coupled system (Kokhanovsky, 2008).Great efforts have been made to study the aerosol from remote sensing data as presented in Kaufman et al. (1997b, a); Holben et al. (1998); King et al. (1999); Levy et al. (2007); Kokhanovsky et al. (2007Kokhanovsky et al. ( , 2010)); Bassani et al. (2010).These studies Figures

Back Close
Full have demonstrated the crucial role of the aerosol optical thickness at 550 nm, τ 550 , in the atmospheric radiative transfer modeling.Several studies (Kotchenova et al., 2008;Vermote et al., 1997b;Kaufman et al., 2002;Guanter et al., 2007Guanter et al., , 2009b;;Gao et al., 2009;Goetz et al., 1985) showed the role of τ 550 on the atmospheric correction of multi-and hyperspectral data dedicated to the retrieval of ocean and land properties.
A series of physically-based algorithms to retrieve aerosol loading (τ 550 ) from remote sensing images have been developed, using MODIS both over land (Kaufman et al., 1997a) over ocean (Tanré et al., 1997) and using the hyperspectral CHRIS-PROBA sensor (Guanter et al., 2005).However, neither the variability of aerosol type nor the possible influence of this variability on the atmospheric correction results is evaluated.
In this context, the focus of this work is to evaluate the impact of aerosol type on the results of the atmospheric correction of the HICO ™ (Hyperspectral Imager for the Coastal Ocean) data acquired in a coastal environment in the northern Adriatic Sea.In a previous study, (Bassani et al., 2012), the aerosol type was connected to the accuracy of aerosol loading for the reflectance retrieval from hyperspectral spaceborne data.In this study a new algorithm, HICO@CRI (HICO ™ ATmospherically Corrected Reflectance Imagery), was developed to perform the atmospheric correction of HICO ™ data.The HICO@CRI is based on the vector version of the Second Simulation of a Satellite Signal in the Solar spectrum (6SV) atmospheric radiative transfer code (Vermote et al., 2006), which is the improved open-source code of the 6S (Vermote et al., 1997b) used for the new generation atmospheric correction algorithm of MODIS data (Vermote and Kotchenova, 2008).
The water reflectance was obtained by running the HICO@CRI algorithm with maritime, continental, and urban aerosol types provided by default by 6SV.An additional run was performed using the micro-physical properties of the aerosol provided by the Acqua Alta Oceanographic Tower (AAOT) station (Mélin et al., 2006) of the widespread AERONET network (Holben et al., 1998).Results from last run allow to build a reference, against which the output obtained by running HICO@CRI with the three 6SV default aerosol types can be compared.The NRMSE (Normalized Root Mean Square) Introduction

Conclusions References
Tables Figures

Back Close
Full was used to evaluate the radiative impact of 6SV default aerosol types with respect to the AERONET inversion products.

Site
The Acqua Alta Oceanographic Tower (AAOT) is located in the northern part of the Adriatic Sea, 15 km off the Venice lagoon (12.51 • E, 45.31 • N).The AAOT (Mélin et al., 2006) is an automated station of the spreadworld AERONET network (Holben et al., 1998).Besides atmospheric acquisitions, the CIMEL sun sky-radiometer instrument at AAOT also performs ocean color measurements for the AERONET-OC network (Zibordi et al., 2009).
The site is in a coastal area generally affected by continental or urban aerosol, as reported in Mélin et al. (2006).Neverthless, since it is located in the Mediterranean Sea, the possible presence of maritime aerosol was also taken into account.

Data
The data used in this study were acquired by using two passive remote sensing instruments operating in the visible spectral domain.The atmospheric data were obtained from the in situ automatic tracking sun sky-radiometer CIMEL CE-318 of the AAOT station (Holben et al., 1998;Smirnov et al., 2011Smirnov et al., , 2009;;Mélin et al., 2006), while image data come from the HICO ™ hyperspectral sensor (Lucke et al., 2011).
The CIMEL CE-318 provides the aerosol optical thickness at the nominal wavelengths (440, 500, 670, 870, 940, 1020 nm), the columnar content of water vapor, wvcc, and ozone, occ, by measuring the direct component of the solar irradiance.From the measurements of the diffuse component on four bands (442, 668, 870, 1020 nm) of the sky radiance at specific angles, the AERONET inversion products, aerosol microphysical and optical properties, are retrieved.Introduction

Conclusions References
Tables Figures

Back Close
Full The atmospheric data have three quality levels: level 1.0 (unscreened), level 1.5 (cloud-screened), and level 2.0 (cloud-screened and quality-assured).The lev 2.0 of atmospheric properties is used for atmospheric correction, and also the level 1.5 has been largely used for remote sensing applications, principally the micro-physical properties being often limited to lev 1.5.
The HICO ™ sensor (Lucke et al., 2011), is the first spaceborne hyperspectral sensor dedicated to the coastal and inland monitoring and ocean characterization.The sensor is composed of 87 contiguous channels in the spectral domain 400-900 nm with a the Full Width at Half Maximum (FWHM) of about 10 nm.The ground swath is about 43.5 km with pixel size of 87 m at nadir acquisition, becoming 93 km with 182 m pixel size at 45 • for the view zenith angle.The signal-to-noise ratio (SNR) is > 200 : 1 over 400-600 nm as reported in Lucke et al. (2011).
Table 1 shows the six HICO ™ images with corresponding AAOT data to constrain the τ 550 , the wvcc, and the occ.The Angstrom coefficient has been also reported for the qualitative evaluation of the aerosol size from the measurements of direct solar irradiance.In order to analyze the contribution of the aerosol type on the water reflectance obtained by using the HICO@CRI algorithm, the AERONET inversion products (microphysical properties) of the AAOT AERONET station and three aerosol types (maritime, continental and urban) were considered for the atmospheric correction.
Figure 1 shows the micro-physical properties, size distribution, and the real and imaginary part of the refractive index retrieved from the diffuse measurements of solar radiance by using the CIMEL CE-318.In the left plot, the bimodal distribution of the aerosol size is in agreement with the results reported in Mélin et al. (2006) non-absorbing particles in the atmosphere.The micro-physical properties, size distribution and refractive index, permit detailed evaluation of the aerosol impact on the atmospheric correction of HICO ™ data.The aerosol optical properties required for radiative modeling and used for the atmospheric correction, were derived from these micro-physical properties by using the subroutines MIE and AEROSOL included in the 6SV source code (Vermote et al., 2006).The 6SV default aerosol types (cf.Sect.2) are defined according to a combination of the four basic components: sea-salt, water-soluble, soot and dust-like (Lenoble, 1985;d'Almeida et al., 1991;Kokhanovsky, 2008;Vermote et al., 2006).Table 2 shows the volumetric percentage of the basic components for the three aerosol types.Figure 2 shows the single-scattering albedo (left) and the asymmetry parameter (right) of the three 6SV default aerosol types and of the measurement days.In Fig. 2 the continous lines represent the optical properties simulated by the MIE and AEROSOL subroutines of 6SV while the dots indicate the corresponding values obtained by the CIMEL measurements.Simulated and measured data are indicated with the same color for each day (cf.Table 1).
It is evident from Fig. 2, the AERONET inversion products are comparable to the spectral behaviour of maritime type, in contrast with the dominating influence of continental and anthropogenic aerosol reported in Mélin et al. (2006).

Methods
The methodology applied in this work is based on the analysis of the contribution of the aerosol type on the water reflectance to improve the accuracy of sea surface reflectance, whose assessment is closely related to the retrieval of water quality parameters (chlorophyll, total suspended matter and colored dissolved organic matter), (Zibordi et al., 2011;Giardino et al., 2010;Guanter et al., 2010;Brando and Dekker, 2003).The physically-based atmospheric correction algorithm, HICO@CRI, which relies on the 6SV code, was developed to accept wvcc, occ, τ 550 and type of the aerosol.Introduction

Conclusions References
Tables Figures

Back Close
Full Concerning the aerosol type, the AERONET inversion products (i.e. the micro-physical properties of the aerosol) or the default aerosol types (i.e.percent of the basic components) can be introduced in the algorithm.
The algorithm for the atmospheric correction of HICO ™ image was implemented in the open source GDL programming language (Coulais et al., 2009;Arabas et al., 2010), following the method reported in Bassani et al. (2010).Starting from the atsensor radiance, L v , the equation presented in Vermote et al. (1997b) was used under the assumption of isotropy surface (Lambertian surface), to neglect the influence of the solar and viewing angles on the surface reflectance, such as the Bi-directional Reflectance Distribution Function (BRDF) of the surface.Thus, the results of HICO@CRI application can be addressed to the aerosol type without contribution of the geometrical conditions to the retrieved surface reflectance.
The equation solved for the surface reflectance (Bassani et al., 2010) for each HICO ™ channel, ρ gnd , is the following: where ρ TOA = πL v /µ s E s is the sensor reflectance expressed by the TOA (Top Of Atmosphere) solar irradiance, E s , and by the cosine of the solar zenith angle, µ s = cos(θ s ).
The residual variables are radiative quantities simulated with a spectral sampling of 2.5 nm covering the domain from 400 nm to 900 nm by using the last version, 6SV, (Kotchenova et al., 2008) of the Second Simulation of a Satellite Signal in the Solar Spectrum, 6S, (Vermote et al., 1997b) radiative transfer code.t g is the gas transmittance; ρ atm is the atmospheric reflectance, called path radiance.S is the atmospheric spherical albedo, (Kokhanovsky, 2008).t S = t ↑ + t ↓ is the total transmittance, obtained from the downward and upward direction.Each one is composed by the direct and diffuse component, t The HICO@CRI algorithm computes the convolution of the radiative quantities, simulated with 2.5 nm spectral sampling, on the spectral response of HICO ™ sensor assumed gaussian with the central wavelength and the FWHM allowed in the header file of the image; procedure is described in Bassani et al. (2010).
The geolocated and geometric input are loaded in the HICO@CRI algorithm from a file supplied with each HICO ™ image, where the following data are contained: longitude (degrees), latitude (degrees), the viewing zenith (VZA) and azimuth (VAA) angles, and the solar zenith (SZA) and azimuth (SAA) angles, referred to each pixel of the image.These values permit to simulate the radiative quantities for each pixel avoiding errors for the calculation of the geometrical configuration of the acquisition.
The surface reflectance, ρ(λ), is finally retrieved removing the effect of the neighboring pixels to the pixel viewing from the sensor, by using the empirical formula employed in atmospheric correction algorithms (Bassani et al., 2010;Guanter et al., 2007Guanter et al., , 2009a;;Kotchenova et al., 2008;Vermote et al., 1997b, a) In Eq. ( 2) ρ gnd is the mean of the pixels adjacent to the viewing pixel covering the swath of the HICO ™ sensor, approximately 30 km 2 .This is the empirical approximation of the environmental contribution to the radiance reaching an airborne or spaceborne sensor.
Concerning the atmospheric input, the HICO@CRI algorithm loads parameters of aerosol and gases.Evaluation of the role of aerosol type on the results of the atmospheric correction over coastal area needs an accurate aerosol loading, by the τ 550 , and the columnar content of the gas with the absorption bands falling into sensor channels.At present, the considered gas parameters are the columnar content of water vapor, wvcc, and ozone, occ.The known and constrained properties related to the gas and to the aerosol loading guarantee that the variability of the atmospheric correction results is only due to the aerosol type.Introduction

Conclusions References
Tables Figures

Back Close
Full In the HICO@CRI algorithm, τ 550 can be varied as well as the aerosol size distribution and the refractive index (if CIMEL CE-318 data have to be used) or the volumetric percentage of the basic components (if default 6SV aerosol types are used).
The aerosol loading is a primary characteristic for the atmospheric radiative transfer applied to the remote sensing data.The optical parameter of aerosol acknowledged to describe the aerosol loading (Kaufman et al., 1997a) is the aerosol optical thickness at 550 nm, τ 550 .Other aerosol optical properties relevant to atmospheric radiative transfer modeling are defined by the aerosol type.Consequently, the radiative quantities of the Eqs.( 1) and ( 2) are expected to depend on aerosol type and thus the results of the atmospheric correction, water reflectance, should be different.
The micro-physical properties, downloaded from the AAOT station and used as input in the HICO@CRI, are the size distribution and the refractive index (cf.Fig. 1).Otherwise, in the developed algorithm, aerosol types can be introduced from the 6SV default types defined by the optical properties (Vermote et al., 2006), described in the Sect.3.

Results
The HICO ™ images, reported in Table 1, were processed by applying HICO@CRI algorithm.The geolocated and geometric inputs were taken from the ancillary data; the atmospheric inputs (τ 550 , wvcc, and occ) were constrained by the AERONET products, while the parameters defining the type of aerosol were varied.In the first step, the atmospheric correction was performed considering the micro-physical properties of the aerosol provided by AAOT site; in the second step, considering the optical properties of the three 6SV default aerosol types (continental, maritime, and urban).Thus, for each HICO ™ image, four corrected images were obtained.
The differences among the retrieved products are analyzed for different water types observed in each HICO image.To this end, three Regions Of Interest (ROIs) were selected on the basis of the variability and representativity of the water surface relatively to each HICO ™ image (cf.Table 1).The Euclidean Distance (ED) between the sensor Introduction

Conclusions References
Tables Figures

Back Close
Full radiance in correspondence of the maximum, 550 nm, and the minimum, 680 nm, of the solar radiation reaching the sensor coming from target of water was used.The radiance measured by the sensor was considered to classify the water surface independently from the atmospheric correction procedure.The locations of the ROIs were defined with similar geometrical configuration to neglect different radiative contribution on the sensor radiance, due to the viewing and illumination angles between the ROIs of each image.Besides, the ROIs were selected close to the AAOT station in order to avoid the impact of horizontal atmospheric variability within the HICO ™ scenes.
Figure 3 shows the ROIs chosen for each HICO ™ image used in this study.The red ROI (ROIrL) is representative of water with low luminosity in the image, the blue and green ROIs represent water with medium (ROIbM), and high (ROIgH) luminosity, respectively.For each HICO ™ image, four mean reflectance values were calculated with respect to the three ROIs obtained for the three 6SV default aerosol types and for the AERONET inversion products.
Figure 4 shows the average of the water reflectance retrieved by using the AERONET inversion products calculated for each ROI (cf.Fig. 3) and used as benchmark for the evaluation of the radiative impact of the aerosol type on the atmospheric correction processing.The retrieval of the water reflectance was performed for high (ROIgH), medium (ROIbM), and low (ROIrL) reflectance of the water.The noisy reflectance can be attributed to the effect of the signal-to-noise (SNR) ratio of HICO ™ sensor calculated for low reflectance cases, as reported by Moses et al. (2012).The negative values of reflectance obtained from correction of the HICO ™ image acquired the 12 May 2012 is explained by the limits of accuracy of the radiative transfer simulation for early morning irradiance case as shown in Table 1.Besides, the negative values of reflectance on 4 May 2011 could be likely due to the loos of accuracy of the simulation results produced by the tri-modal size distribution obtained from CIMEL measurements, as shown in Fig. 1.
Figure 5 shows the ED between the reflectance retrieved using the AERONET inversion products and the 6SV default aerosol types (maritime, continental, and urban).Introduction

Conclusions References
Tables Figures

Back Close
Full The ED is reported for the HICO ™ channels centered in atmospheric windows, to point out the aerosol radiative impact avoiding the gas influence during the extinction modelling.Panels on the left show the ED for the image with the minimum of the τ 550 = 0.01 and panels on the right shows the results for the maximum, τ 550 = 0.27.In order to quantify the contribution of each 6SV default aerosol type on the spectral water reflectance with respect to the AERONET inversion products, the Root Mean Square Error (RMSE) was normalized to the distance between the outer limits (maximum and minimum) of each spectral water reflectance.
Figure 6 shows the Normalized Root Mean Square Error (NRMSE) vs. the τ 550 for the three ROIs, where the water reflectance accuracy decreases with growing aerosol loading, particularly when an unsuitable aerosol type is selected.The aerosol loading changes slowly the accuracy of atmospheric correction results if an aerosol type, having similar optical properties of the AERONET inversion products, is considered.In the case study, the aerosol type which better satisfies the condition is the maritime one, while, if a continental aerosol is assumed, the accuracy decreases less if urban is selected.
During acquisition periods, aerosol was mainly distribuited in the fine mode, showing a non negligible imaginary part of refractive index, with the exception of 4 May 2011 and 25 April 2013.The NRMSE for the 4 May 2011 is higher than for the other ones (Fig. 6).The imaginary part of the refractive index is negligible only for the aerosol performed by using the urban aerosol in the HICO@CRI algorithm.The water reflectance is mainly correlated for the AERONET inversion products and the maritime type, independently from the luminosity of the ROIs, like attested by Fig. 6.This is observed also for the 4 May 2011 case, although the water reflectance obtained from the HICO@CRI with AERONET inversion products for the ROIrL shows negative values.
Furthermore, the tri-modal size distribution observed on 4 May 2011 could be related to the spike shown in the Fig. 6 for τ 550 = 0.08.This result suggests that the sensitivity of the 6SV to a tri-modal size distribution needs to be investigated more in depth when atmospheric correction of hyperspectral data over water surface is performed.Finally, errors on the retrieved water reflectance are greater for the continental and urban aerosol types than for maritime one.

Discussion and conclusions
The study shows the impact of different aerosol types on the results of the physicallybased atmospheric correction of hyperspectral data characterized by low reflectance.
To this aim, the HICO@CRI (HICO ™ ATmospherically Corrected Reflectance Image) atmospheric correction algorithm was developed and applied to HICO ™ images acquired on the northern Mediterranean Basin, when the inversion products of the AAOT (Acqua Alta Oceanographic Tower) site were available.The aerosol loading defined by the aerosol optical thickness at 550 nm, τ 550 , and the columnar content of the water vapor, wvcc, and ozone, occ, were constrained to the values downloaded from the AAOT site.The analysis has been performed for the HICO ™ images acquired for different values of the aerosol loading (τ 550 < 0.3) and with clear-sky condition.A measured aerosol, described by size distribution and refractive index, obtained at AAOT station was used as benchmark to evaluate the accuracy of the water reflectance obtained from the atmospheric correction performed using the maritime, continental and urban 6SV default aerosol types.Results show that the accuracy of the retrieved water

AMTD Introduction Conclusions References
Tables Figures

Back Close
Full reflectance depends on the aerosol type used in the atmospheric correction processing, in addition to the influence of aerosol loading.The Normalized Root Mean Square Error (NRMSE) attests the decrease of the retrieved water reflectance accuracy with growing aerosol loading.The accuracy of the atmospheric correction results depends on the τ 550 and also on the aerosol type, mainly if the aerosol loading is non negligible.
In the case study, the AERONET micro-physical properties suggest that the aerosol type during the HICO ™ acquisition is comparable to the maritime 6SV default aerosol type.The NRMSE increases if an unsuitable aerosol type is considered during the atmospheric correction, principally if the aerosol is composed by basic components with non negligible absorption, like the urban type.Thus, the improvement of the water reflectance retrieval requires more attention on the aerosol type during the atmospheric correction processing.The presence of absorbing aerosol leads to a significant contribution of the absorption on the radiative simulation of the solar radiation extinction and, consequently, in the water reflectance obtained from the Eqs.( 1) and ( 2).The absorption of the urban type is non negligible; consequently the extinction is not dominated by the scattering, like for the maritime and continental types.This situation is confirmed by the optical properties of the urban aerosol presented in the Fig. 2.
Results of this work highlight the dependence of the atmospheric correction of water surface on the aerosol type.The sensitivity to the aerosol type derives from the components of the radiative field in the Eqs.( 1) and ( 2), simulated for specific values of micro-physical and optical properties of the aerosol, by using the 6SV radiative transfer code (Kotchenova et al., 2008).During the HICO ™ acquisition considered in this study, the Northern Mediterranean Basin was characterized with maritime aerosol, whereas the continental and urban types seem to be not consistent with the AERONET inversion products and to lead to inaccurate atmospheric correction results.Consequently, the radiative impact of the aerosol type is not negligible when the atmospheric correction of hyperspectral data over coastal environment is performed.The retrieved water reflectance is sensitive to the aerosol type and the suitable aerosol

Conclusions References
Tables Figures

Back Close
Full type has to be chosen to avoid errors on the water reflectance evaluation.The results of this work might be useful for applying atmospheric correction algorithms to remote sensing data used in water quality applications.

Conclusions References
Tables Figures

Back Close
Full  Full Starting from the left, the calculation for the low, medium and high reflectance water types.
Results of these case studies highlight how the water reflectance accuracy decreases for higher aerosol loading, particularly, when the atmospheric correction is performed by using the urban aerosol in the HICO@CRI algorithm.The water reflectance is mainly correlated for the AERONET inversion products and the maritime type, independently from the luminosity 10 of the ROIs, like attested by Figure 6.This is observed also for the 2011/05/04 case, although the water reflectance obtained from the HICO@CRI with AERONET inversion products for the ROIrL shows negative values.Furthermore, the tri-modal size distribution observed on 2011/05/04 could be related to the spike shown in the Figure 6 for τ 550 = 0.08.This result suggests that the sensitivity of the 6SV to a tri-modal size distribution needs to be investigated more Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | , except for the 25 April 2013 where the accumulation mode prevails in the volume size distribution.The 4 May 2011 presents a tri-modal size distribution, observed for low aerosol loading in the site (Mélin et al., 2006).The other plots show the real and imaginary parts of the refractive index.The spectral values of the imaginary part attest non negligible absorption of aerosol in the coastal area during data acquisition, especially on 11 January 2012, while observations on 4 May 2011 underlines the presence of Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | s ) are the cosine of the view (v) and sun (s) zenith angles and τ is the total optical thickness due to the aerosol and gases.Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | type during the HICO ™ acquisition of the 4 May 2011, as shown in Fig. 1.Thus, the radiative impact of aerosol absorption plays a crucial role on the atmospheric correction of hyperspectral data over water target in coastal environment.On the other hand, the Fig. 1 shows a size distribution of the 25 April 2013 with a predominance in coarse mode.The aerosol size affects weakly the accuracy of the retrieved water reflectance, as shown in Fig. 6 in the case study of 25 April 2013 defined with high aerosol loading, τ 550 = 0.27.Results of these case studies highlight how the water reflectance accuracy decreases for higher aerosol loading, particularly, when the atmospheric correction is Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Figure 1 .Figure 2 .Figure 3 .Figure 5 .Fig. 6 .
Figure 1.An example of the micro-physical properties of the aerosol downloaded from the AERONET AAOT station.The micro-physical properties values are the size distribution (on the left), the real part of the refractive index (on the centre), and the imaginary part (on the right), of the CIMEL acquisition.

Figure 6 .
Figure6.The Normalized Root Mean Square of the water reflectance retrieved with the aerosol types (urban, continental, and maritime) respect to the AERONET inversion products, calculated for the ROIs.Starting from the left, the calculation for the low, medium and high reflectance water types.

Table 1 .
The simultaneous HICO ™ and CIMEL acquisition with the τ 550 , wvcc, occ and the Angstrom coefficient downloaded from the AAOT AERONET station.The selected images meet the clear-sky requirement.