Current research:
Recently completed research:
Validation of GOME and SCIAMACHY Ozone measurements (EO-039)
PI: D.P.J. Swart, RIVM
This project aims at validating ozone profile measurements made from satellites. The emphasis is on ozone profiles from GOME (on board ERS-2) and from GOMOS, MIPAS and SCIAMACHY (on board Envisat). As a reference, high quality ground-based measurements are used which are taken within the NDSC (Network for the Detection of Stratospheric Change). Among these, a special role is reserved for the RIVM stratospheric lidar in Lauder, New Zealand. This instrument measures the ozone profile using pulsed laser light, and is the only routine monitoring instrument of this quality on the Southern Hemisphere.
In 2003, our first validation results for GOME (processed with a KNMI algorithm) were published, showing a useful data range from 17 to 50 kilometers altitude. Within this range, differences remain which vary with season and altitude. In addition, tools and guidelines were developed for non-expert satellite data users. In ESA framework, this analysis has been extended to include the major competing GOME algorithms worldwide. These analyses are now well under way, and results will be published in 2004.
For Envisat in 2003, the focus was on GOMOS ozone profiles. This choice was mainly driven by satellite data availability. A large worldwide dataset has been obtained of paired observations by ozone sondes, groundbased lidars, ozone microwave instruments and GOMOS, using well-defined collocation criteria. Preliminary results show an excellent performance of this instrument in the altitude range of 18 to at least 60 kilometers altitude. In 2004, this work will be finalized and published.
Monte Carlo radiative transfer in spherical geometry applied to the measurements made by SCIAMACHY (McSCIA) (EO-041)
PI: Dr M. Krol, IMAU
SCIAMACHY, launched on 1 March 2002, measures the radiation that is scattered back by our atmosphere. From this backscattered radiation it is possible to determine concentrations of gases in the atmosphere, e.g. ozone, NO2, formaldehyde, etc. The retrieval of trace gas information from backscattered solar radiation is currently an active research theme. SCIAMACHY has several modes of viewing the Earth. In the nadir viewing mode the radiation perpendicular to the Earth’s surface is measured. In the limb mode, the satellite looks through the atmosphere. In order to obtain valuable information from the backscattered solar radiation, radiative transfer models are needed. For the nadir observing mode, plane parallel radiative transfer models can be used. However, for the limb mode this is not possible. A way to model the radiative transfer for limb observations is by means of Monte Carlo radiative transfer simulations. The photons of the sun are traced through the atmosphere, and the physical processes, like absorption and scattering, are modelled as a stochastic process. The McSCIA project aims at the development of such a Monte Carlo radiative transfer model, in order to analyse the SCIAMACHY limb and nadir measurements. After the validation against the state of the art Radiative Transfer Models, McSCIA was used to study 3D features of the UV-vis-Nir solar radiation. A simple case study has shown the need for more investigation on the relation between the satellites increasing horizontal spatial resolution and the real area seen by the solar radiation. This could affect the results of satellite measurements and the spatial location attributed to some trace gases like Ozone and NO2.
The determination of geopotential heights at constant pressure surfaces using Radio Occultation and application to climate change detection (EO-043)
PI: Dr S.J.M. Barlag, KNMI
The main aim of the project is the creation of geopotential height maps of pressure surfaces from Radio Occultation data obtained from various sources. These maps can be used to monitor the development of average atmospheric temperatures with time, since these temperatures are directly connected to geopotential heights of pressure surfaces. These temperature records at altitude can thus be used in synergy with surface temperature data.
Tangible results so far include:
An accurate operational water column for GOME and SCIAMACHY: improved database and data product (EO-046)
PI: Prof Dr W.J. van der Zande, FOM-AMOLF
An efficient global mapping of trace gases and aerosols are prerequisites for studies into their long-term impact. Sensitive satellite instruments, such as GOME and SCIAMACHY, are needed, as well as a thorough knowledge of radiation transport, and light scattering properties. A high computational efficiency is needed in view of the data streams involved.
An efficient computational scheme was introduced, which has made it possible to treat spectra of trace gases containing tens of thousands of individual absorption lines. This method is now adopted at the Max Planck Institute for Air Chemistry, Mainz for long-term studies of water vapour. Our attempts to compare and improve upon water vapour data bases have found a place in international collaborations; satellite data and the atmosphere act as our laboratory. At the Vrije Universiteit, first steps have been made to validate an effective laser-based absorption measurement technique; cavity ring down spectroscopy (CRDS) for the determination of a data base. This project will be continued at the University of Nijmegen. The extension of our methods from global maps of water vapour columns to global maps of altitude distributions of the water vapour (water profiles) is carried over to the new position of the graduate student Rüdiger Lang.
In the visible and near infrared part of the spectrum clouds as well as aerosols complicate attempts for quantitative determinations of trace gases. Solutions have to be found in instrumental developments, smaller ground pixels for less interference by clouds, and by observing for example the polarization of the scattered sunlight. Solutions also have to be found in algorithm developments. A lot of work has been put by our group in the understanding the impact of mineral aerosols on light scattering.
In the coming final year, we hope to contribute to most of these topics. We hope to generate a database for the description of the light scattering properties of mineral aerosol. This database should improve the identification of aerosol by remote sensing. Furthermore, we hope to perform field measurements with a high resolution instrument searching for line shape properties of water vapour.
Retrieval of ice and water cloud properties from SCIAMACHY's near-IR channels (SCIA-CIRRUS) (EO-047)
PI: Dr P. Stammes, KNMI
The main objective of this project was to derive global distributions of ice clouds (cirrus) from near-infrared spectra of the Earth measured by SCIAMACHY. From the refractive index spectrum of water and ice, we knew beforehand that there is a strong difference between the absorption spectra of water and ice in the near-infrared around 1600 nm. Since SCIAMACHY is uniquely equipped to measure spectra of clouds in the near-infrared (contiguous spectra up to 1750 nm), this project was aimed at obtaining cirrus cloud information from spectral discrimination. In order to reach our goal two preparatory steps had to be taken: (1) radiative transfer modelling of ice clouds, and (2) investigation of the calibration of SCIAMACHY. After these steps were completed, the goal of making global maps of ice clouds from SCIAMACHY could be reached. This work led to a new SCIAMACHY product, called the Cloud Phase Index (CPI). This product can be used to discriminate water from ice clouds. In the three steps of the project the following results were obtained.
(1) Radiative transfer modelling of ice clouds is complicated, because of the non-spherical shape of ice crystals. A-priori knowledge of the ice crystal shape is needed for the global retrievals of cloud optical thickness and particle size. For water clouds, the sphere is an adequate particle shape model, but for ice clouds more complex shapes are necessary for accurate retrievals. In order to find a suitable ice crystal model for global retrievals, we tried out different ice crystal models (perfect hexagons, inhomogeneous hexagons, imperfect hexagons, and fractal particles) for the evaluation of global POLDER multidirectional total and polarized reflectance measurements. We found that imperfect or inhomogeneous crystal shapes allow for a much better representation of global POLDER reflectance measurements than the perfect hexagonal shape. The imperfect hexagonal model was therefore used in the next steps.
(2) In the first period after the launch of SCIAMACHY much work was needed to investigate the quality of the level-1 data, which are the original radiation measurements. The main level-1 quantity is the Earth reflectance. For all cloud retrievals the quality of the reflectance calibration is essential. It readily appeared that SCIAMACHY is underestimating the reflectance, but the question was: how much precisely? By comparing data from SCIAMACHY with data from MERIS, which is also onboard Envisat, we found that SCIAMACHY has a reflectance error of up to -20 %. A simple correction for this error is possible.
(3) The expected difference in the shape of the reflectance spectrum of water and ice clouds around 1.67 micron has indeed been found in cloud observations by SCIAMACHY. The spectral slope around 1.67 micron has been quantitified as the so-called Cloud Phase Index. This CPI was validated with colocated MODIS measurements of cirrus reflectance. Global maps of the CPI have been presented, indicating the presence of cirrus clouds around the Equatorial region. Furthermore, regions covered by snow / ice in the Northern hemisphere can also be seen in the CPI. Hence, there is a sensitivity to the ground type. Future work includes the combination of the developed CPI with other cloud detection algorithms, like the much used FRESCO algorithm. This will further improve the characterization of clouds from SCIAMACHY.
Resolving spatial and temporal atmospheric water vapour structures using a ground based GPS receiver network (EO-050)
PI: Dr S.J.M. Barlag, KNMI
The subject of this project is the retrieval of three-dimensional atmospheric water vapour distributions from atmospheric signal delay values provided by geodetic networks using the Global Positioning System (GPS).
GPS Zenith Total Delay has correlated errors even over large distances. This is due to the fact that in order to estimate several unknowns (e.g. positions of the GPS receiver) all observations in a certain time window from a number of sites are used to determine the delays at these sites. This introduces correlated errors in ZTD. This correlation will also be present in Slant Delay observation.
Different methods are developed to obtain uncorrelated (zenith and slant) observations. A tool is developed to reconstruct 3D water vapour from slant observations.
Sensor synergy algorithms for CloudSat and PICASSO-CENA (EO-052)
PI: Dr D.P. Donovan, KNMI
This project will contribute to improving the understanding of clouds and their treatment in atmospheric models. Through their interaction with solar and thermal radiation clouds are extremely important in determining atmospheric and surface heating and cooling rates. Cloud processes are very complex and in general are not being adequately treated in atmospheric models, especially as there is no vertical cloud-profile information going into these models. The specific goal of this project is to develop a procedure for combining a space-based lidar (laser-radar) together with simultaneous cloud-radar measurements in order to measure vertical profiles of cloud properties on a global base. From these observations the statistical (vertical) properties of ice-clouds will be derived and inserted in atmospheric models.
The last two years the procedure was tested on ground-based data. A comparison was made between ice cloud properties at sites in the Netherlands, England and the USA. In various atmospheric models ice cloud effective particle size is often assumed to be temperature dependent. It was found that the relationship between cloud effective particle size and temperature was different between the European sites and the American site in the central U.S. However, the ice particle sizes were found to be correlated to the depth into the cloud from cloud top resulting in a single parameterization valid at the three sites and possibly globally. The consequences of this finding are currently being tested in a regional climate model (RACMO) at KNMI.
Validation and interpretation of SCIAMACHY's polarisation measurements (SCIA-POLARISATION) (EO-054)
PI: Dr P. Stammes, KNMI
The aim of this work is to contribute to the calibration of SCIAMACHY. The specific aim here is to validate the polarisation correction of the radiance spectra. This correction is important for the quality of a large number of scientific products from SCIAMACHY, like ozone profiles, aerosol and cloud properties.
In 2004 an alternative polarisation retrieval algorithm was set up, with which we were able to validate the polarisation product of the operational Level-1 product. As of software version 5.00, we have a very good agreement with the most important PMD in nadir mode, PMD 1. The improvement in going from software version 4.02 to 5.00 was large, and shows that it is feasible to get all PMDs working properly in the future, for nadir and limb mode.
However, despite all the improvements achieved in 2004, there will still remain a lot of problems with the data in 2005, and the Level-1 validation work is not yet finished. We will therefore continue to improve the polarisation correction algorithm and the validation of the SCIAMACHY polarisation data.
Radar - lidar synergy for space-based retrievals of water cloud parameters (EO-060)
PI: Dr H.W.J. Russchenberg, DUT
Knowledge of water clouds is very important for climate studies, but retrieval of their micro-physical properties from space is difficult and challenges the technological possibilities of today. This is partly due to the limitations of space-based instruments and partly due to physical processes in water clouds that obscure direct relationships between remote sensing observables and the cloud parameters. In an earlier project at DUT/IRCTR, a new technique was developed that overcomes these limitations by means of radar and lidar synergy. It is the objective of this project to further develop the technique and adapt it to the specifications of the future EarthCare mission.
In preparation for the future atmospheric remote sensing satellite missions (CloudSat-Calipso and EarthCare) a big volume of ground-based observations was analysed and selected to test the radar-lidar water cloud parameters retrieval algorithm. After software development these data were processed and a database of retrieval results that continuously covers the time period after October 2002 for two ground-base sites – Cabauw (The Netherlands) and Chilbolton (UK) was created. Using this database and the available aircraft dataset (CLARE’98 campaign) the experiments with varying range resolution of the radar, data analysis of the impact of range resolution on retrieval techniques have been done and their results show that developed algorithm is applicable for the processing of the CloudSat/Calipso data. A more detailed study directed to the improvement of the retrieval technique will be done next year. After the launches of the CloudSat and Calipso satellites, which are scheduled for summer 2005, their data will be included in the analysis.
PI: Prof Dr W. Ubachs, VU
In the laboratory experiments performed at the Laser Centre Vrije Universiteit, the cavity-ring down laser technique is employed to determine cross sections for the O2-O2 collisionally-induced absorption resonances. With the use of the ´pressure-ramp method´ the contribution of Rayleigh scattering (linear in the pressure) can be disentangled from true collisional effect (quadratic in pressure). In the past year measurements have been performed for the 477 nm and 577 nm resonances at temperatures in the range 184 – 294 K, corresponding to the atmospherically relevant range. Important findings are that the widths of the resonances increases with pressure, in particular for the blue 477 nm resonance, where a linear increase of the width is found in the range 200 – 300 K from 150 to 260 cm-1. The peak cross section is found to be linear, within error limits, over the entire atmopheric window 200 - 300 K. The band integrated cross section is found to slightly increase over this temperature range. These results, that have been submitted for publication, bear relevance for cloud retrieval schemes in atmospheric Earth observation, in particular for the Ozone Monotoring Instrument (OMI).
OPAES: Optimal estimation of primary aerosol emissions from satellites (EO-063)
PI: Prof Dr A.W. Heemink, DUT
Atmospheric aerosols have a large impact on climate, human health, visibility and continental and maritime ecosystems. Whereas formation and distribution of secondary inorganic aerosols is nowadays relatively well understood and can be modelled quite accurately, the understanding of the primary emissions is still very incomplete. Satellite retrieved aerosol fields contain a lot of information on aerosol concentrations and have a large added value over the aerosol data from the many European monitoring networks which provide point measurements that are often concentrated in certain regions.
Data assimilation can be used to combine the results of a numerical atmospheric chemistry model with the measurement information available in order to obtain an optimal reconstruction of the dynamic behaviour of the aerosol concentrations. Most existing data assimilation schemes are developed and used for prediction problems. In the first phase of the project a data assimilation scheme will be developed especially designed for emission reconstructing problems.
Validation of SCIAMACHY and OMI NO2 and aerosol data using Dutch ground-based measurements (EO-065)
PI: Dr P.F. Levelt, KNMI
The Ozone Monitoring Instrument (OMI) was successfully launched on July 15, 2004, initial OMI retrievals look promising. OMI and SCIAMACHY nitrogen dioxide (NO2) and OMI aerosol data will lead to better understanding of the sources and transport of pollution. We describe the progress in the DANDELIONS project, which focuses on aerosol and NO2 satellite validation and algorithm improvement.
The vertical distribution of NO2 in the troposphere influences the environmental impact and the accuracy of satellite retrievals. RIVM is the first to use lidar NO2 measurements to study this distribution for satellite validation and algorithm development. First measurements in the bottom 2 km of the atmosphere are successful. We plan to extend this range, perform more measurements, and study spatial variability of tropospheric NO2.
Aerosols are generated near the Earth’s surface by a variety of processes. They show large temporal, spatial and vertical variations, which may strongly influence satellite retrievals of aerosol. The distribution influences health effects, since these are caused by ground-level aerosols only.
The retrieval algorithms have been prepared for utilization of AATSR on ENVISAT, which has high spatial resolution (1x1 km2 nadir) and is therefore suitable to determine aerosol field variations. Further, we prepared for aerosol monitoring, including preparation of instrumentation for aerosol microphysical and chemical properties measurements.
OMI and SCIAMACHY validation is the focus of a campaign that will be organised in Cabauw in May/June 2005. All instrumentation from our project, along with NO2 instrumentation and models from international partners, will participate.
PI: Dr B. Bregman, KNMI
The aim of the research is to study the interaction between increasing greenhouse gases and air quality. For this purpose two different global models will be used and configured. These models are a chemistry-transport model with detailed chemistry processes. The second model is a climate model. The aim is to couple these two models, and a test version of the chemistry-transport model has already been coupled to the climate model. In addition, the chemistry-transport model is being validated and tested. A code has been developed in this project to compare the model results with satellite data, and has been explored by using GOME ozone profile data.
In the next period the full-chemistry version of the chemistry-transport model will be tested and validated with satellite data. The aim is then to couple this version to the climate model for evaluation of the chemistry-climate interactions.
FRESCO+: an improved O2 A-band algorithm for retrieval of cloud properties from GOME and SCIAMACHY (EO-067)
PI: Dr P. Stammes, KNMI
Clouds strongly influence retrievals of tropospheric trace gases and aerosols from GOME and SCIAMACHY. Accurate co-located cloud information, such as cloud fraction, cloud pressure, and cloud optical thickness are needed to improve tropospheric retrievals from GOME and SCIAMACHY. The O2 A-band from 759-771 nm is well suited to provide this information. Since O2 is a well-mixed gas, the measured column amount of O2 yields the cloud pressure. In the past years the simple but efficient FRESCO algorithm (Fast Retrieval Scheme for Clouds from the O2 A-band) was developed, which is used by several GOME data user groups. The FRESCO algorithm uses the O2 A-band reflectances, and produces effective cloud fraction and cloud pressure with about 50-100 hPa accuracy. In FRESCO, clouds are assumed to be Lambertian reflectors. This limits the accuracy and yields a bias in the cloud pressure results. Therefore, in this FRESCO+ project it is proposed to improve the accuracy of the method by including O2 absorption inside the cloud and allowing for molecular scattering in the clear and cloudy atmosphere. This will be done by extending our radiative transfer model DAK (Doubling-Adding KNMI) with a modified k-distribution approach. In addition to more accurate cloud pressure and cloud fraction products, an extra cloud parameter will be retrieved, namely the O2 absorption path inside the cloud, which is related to the cloud optical thickness and cloud geometrical thickness. Validation will be performed by intercomparison between SCIAMACHY cloud pressures and colocated cloud information from the imagers MERIS and AATSR, which are also on board Envisat. The results from the new algorithm will be made available to the GOME and SCIAMACHY users’ community through, among others, the TEMIS Data Users Programme project of ESA.
SCIALINA: Ozone profile retrieval from combined limb and nadir observations from SCIAMACHY (EO-068)
PI: Dr R.F. van Oss, KNMI
The SCIAMACHY instrument on board the ENVISAT satellite has the unprecedented capability to observe the Earth’s atmosphere in two different ways: looking straight down (nadir mode) and observing the horizon (limb mode). The aim of these observations is to derive information of the distribution of atmospheric chemical constituents, such as ozone.
The Royal Netherlands Meteorological Institute KNMI has developed a computer code that can determine the vertical distribution of ozone and other species from SCIAMACHY limb observations. For this it is necessary to simulate the propagation of solar light through a realistically modelled Earth atmosphere. By comparing the model output with the satellite measurements, the true composition of the atmosphere can be derived in an iterative manner.
Theoretical studies have shown that the combination of nadir and limb observations result in significantly improved vertical profiles of atmospheric trace gasses. Therefore, the developed computer algorithm is currently being extended as to combine both limb and nadir observations in the retrieval process. Thereby, the nadir observations will provide a first estimate for the distribution of, for example, ozone. This way, a number of nadir retrievals can account for the variation of the ozone concentration along the line-of-sight of the limb observations. Subsequently, the limb measurements are applied to improve the derived vertical ozone concentration profiles.
Retrieval of aerosol properties from satellite data (EO-069)
PI: Dr J. Landgraf, SRON
Aerosols are believed to directly affect the Earth's climate by interaction with solar and terrestrial radiation, and indirectly by changing the properties of clouds. The total effect of aerosols represents one of the largest uncertainties in climate research, because its magnitude and even its sign are unknown. In order to better understand the effect of aerosols on climate satellite measurements play a crucial role. The aim of this project is to develop a new algorithm for the retrieval of aerosol properties from satellite measurements of intensity and polarization. The retrieval method is based on an accurate radiation transfer model for polarized light and an analytical inversion approach. The advantage of this new method, compared to earlier methods, is that less ad hoc assumptions about aerosol microphysical properties need to be made and that it allows a solid error analysis and quantification of the information content of satellite measurements. At this stage of the project the main parts of the retrieval algorithm (radiation transfer and inversion model) have been completed and the algorithm has been tested for synthetic intensity and polarization measurements of the GOME-2 satellite instrument, to be launched end 2005. In the next period the algorithm will be adjusted so that it can be applied to real satellite measurements. Furthermore, different measurement concepts for aerosol retrieval will be compared and recommendations will be made for a future satellite instrument suited for aerosol retrieval.
Ozone Monitoring Instrument cloud product validation and interpretation (OMI-Clouds) (EO-072)
PI: Dr P. Stammes, KNMI
The Ozone Monitoring Instrument (OMI), to be launched on EOS-Aura in 2004, will measure the daily global distribution of O3, NO2, BrO, SO2, and HCHO with unprecedented spatial resolution (pixel size of 13x24 km2 subsatellite). To detect the total column amount of these (tropospheric) trace gases accurately, the cloud pressure and effective cloud fraction in the same pixel are needed. This cloud information is retrieved operationally with a new algorithm using the O2-O2 absorption band at 477 nm (O2-O2 is a collision complex of oxygen).
The O2-O2 algorithm uses a DOAS (differential optical absorption spectroscopy) approach of fitting the atmospheric reflectance spectrum to the absorption cross-section spectrum of O2-O2, where the continuum reflectance is used to determine the effective (radiance-equivalent) cloud fraction.
In this project we propose firstly to verify and validate the effective cloud fraction and cloud pressure derived from the O2-O2 algorithm by comparison with colocated groundbased radar/lidar observations of cloud bottom and cloud top in Cabauw (the Netherlands), and with satellite-based observations of cloud fraction and cloud pressure from the Moderate Resolution Imaging Spectroradiometer (MODIS) on EOS-Aqua. In addition we will also compare with the alternative OMI cloud pressure product based on rotational Raman scattering (Ring effect) in the ultraviolet. Secondly, to better interpret the results of these comparisons, detailed sensitivity studies will be performed for multi-layer cloud systems and snow/ice covered surfaces.
Synthesis of active and passive measurements for atmospheric model improvement (Synthesis) (EO-073)
PI: Dr A. Feijt, KNMI
Advances satellite instruments such as MSG and MODIS enable the retrieval of atmospheric physical properties at high resolution. This enables a wide range of applications in climate research and the weatherservice. Thus the use of satellite data is expected to increase dramatically.
However, atmospheric processes show variability on a wide range of scales, from micrometer radiative cooling of cloud water droplets to large scale lifting in a frontal zone at a scale of thousands of kilometers. This implies that observing systems measure a sample of the atmosphere that is only representative for a very short time and very small area. Therefor, there has to be developed a method to correlate measurements of different time and/or spatial scales before the new measurements from MSG can be used in full. The cloud variability issue effects progress in:
This study aims to develop a method for automatic determination of the probability that two samples from ground and satellite relate to the same cloud field. The optimum area and time period will be selected. The relationship between the two sets are defined and thus the comparison (validation) can be made.
The high end product of this study is a method to generate a three dimensional field of cloud properties at the scale of a climate model (50x50km) from the synthesis of time series of detailed ground based remote sensing measurements and satellite derived spatial distribution of vertical integrated cloud properties.
PI: Prof Dr G. de Leeuw, TNO-FEL
The objectives of the project are:
1. to improve the accuracy of the AOD retrieved by using the single and dual view ATSR-2/AATSR algorithms
2. to extend their applicability to the global scale by:
3. The use of the algorithm in a synergistic approach to retrieve the AOD from MSG/SEVIRI data thus providing AOD fields with high spatial and temporal resolution.
The results will provide the scientific input for the development of tools for the assessment of aerosol effects on radiative forcing, air quality and ecosystems. Users are involved in the project as partners to ensure the development of data products meeting their requirements.
Combined active and passive cloud remote sensing using CLOUDSAT and CALYPSO (EO-083, new project)
PI: Dr D. Donovan, KNMI
This project will contribute to the process of improving the treatment of clouds in atmospheric models. Through their interaction with solar and thermal radiation, cirrus clouds composed of ice particles are important in determining atmospheric and surface heating rates. Due to uncertainties concerning the structure of ice clouds and the interactions between microphysics and radiation in these clouds, they are not well treated in atmospheric models. Current space-based observations are unable to provide height resolved data on cirrus clouds critical to improving our understanding of the role of cirrus clouds in climate. Active instruments (cloud radars and lidars) as opposed to traditional passive cloud sensors make direct height-resolved measurements.
This proposal will advance the development of multi-sensor cloud remote sensing techniques and apply such techniques to the study of cirrus clouds. This proposal builds on a combined radar and lidar (laser-radar) approach developed at KNMI as well as a work carried out under a current SRON project. The current SRON funded activity involves adapting an existing lidar/radar approach to data from the CLOUDSAT (radar) and CALIPSO (lidar) missions which will be launched in April 2005 and will fly in a tight formation. This proposal will provide 3-D information on cirrus cloud structure by combining the lidar and radar data with more traditional passive sensors carried on-board CALIPSO. Data from CLOUDSAT/CALIPSO will then be analysed with the aim of assessing, on a global scale, cirrus cloud properties relevant to atmospheric radiative transfer. This information can be used to improve climate and forecast models.
Tropospheric Ozone Re-Analysis (TORA) (EO-084, new project)
PI: Dr M. van Weele, KNMI
The goal of the TORA project is to create a global tropospheric ozone climatology based on measurements by the GOME instrument, for the period 1996-2001. It is estimated that the accuracy of the climatology will be sufficiently high to make it valuable for the study of the inter-annual and seasonal variability and the regional differences in tropospheric ozone and related radiative forcing.
The climatology will be created by substracting the stratospheric part from the total ozone column, where the stratospheric part is obtained from assimilation of GOME ozone profiles into an atmospheric chemistry model. Theoretical error estimates will be provided. Ozone sondes will be used for the validation of the absolute columns and the surface-to-tropopause mean mixing ratios and associated uncertainties.
High-resolution 2D water vapour field estimation by permanent GPS networks in combination with Envisat-MERIS data (EO-085, new project)
PI: Dr Ir H. van der Marel, DUT
Water vapour is the most important, but least understood greenhouse gas on earth. Mapping the spatial distribution of water vapour in the earth's atmosphere is difficult due to the limited spatial and temporal resolutions of contemporary meteorological instrumentation. The main objective of this proposal is to combine observations from ground based GPS and the MERIS instrument on-board Envisat to improve spatio-temporal water vapour distribution retrievals, exploiting the high accuracy, reliability and temporal resolution of GPS, and the high spatial resolution of MERIS. Permanent GPS networks provide high-accurate water vapour observations at the station positions with high temporal resolution. The inter-station spacing is nevertheless large, and simple interpolation between the stations is limited in quality. The MERIS instrument on Envisat retrieves integrated water vapour by observing the backscatter of solar radiation in the near infrared over land, sea and above clouds. With a maximum spatial resolution of 300 m, MERIS can observe dynamic structures on scales much smaller than possible before. Each technique has its own limitations, systematic errors and problems, but by combining these data and comparing it with other data we hope to gain valuable insight in both systems. Our main goal is to develop an application that would give a wide-scale, fine-resolution, total vertical column water vapour field, using GPS, MERIS and wind-field data. Improved knowledge of atmospheric water vapour and its temporal and spatial variability is not only of great scientific interest for climate research and weather prediction, but also will benefit geodetic positioning applications using GPS and radar interferometry.
Use of rotational Raman Scattering to improve Ozone Profile retrieval Algorithms for nadir pointing spectrometers (URSOPA) (EO-086, new project)
PI: Dr J.F. de Haan, KNMI
Various ozone profile retrieval algorithms for spectrometers, like GOME, SCIAMACHY and OMI, have been developed the last decade. However, generally rotational Raman scattering (RRS) is not accurately taken into account. Recently, it was found by various research groups that properly accounting for RRS is essential for differential optical absorption spectroscopy (DOAS) based total ozone column algorithms. Although, the ozone profile algorithm is not based on DOAS it also has to account for the differential absorption structures of ozone. As RRS removes in part these differential structures, properly accounting for this effect will improve the ozone profile retrieval algorithm. From a different viewpoint, there are two streams of light emerging from the atmosphere, one due to elastically (Cabannes) scattered light and one due to inelastically (Raman) scattered light. These streams have a different spectral fine structure and can be distinguished. Currently, the fine structure of the Raman stream is (inaccurately) removed by fitting a differential Ring spectrum, and the remaining part is treated as Rayleigh scattering which is then used to determine the ozone profile. However, extending the algorithm by using both streams will improve the accuracy of the retrieved ozone profile. It is anticipated that significantly more information on tropospheric ozone can be obtained in this manner, which is important for measuring tropospheric pollution and stratosphere-troposphere exchange. The objectives of this project are a) to test and improve an extended ozone profile retrieval algorithm which takes RRS properly into account; b) to perform sensitivity studies to determine nature and magnitude of the improvements; and c) to apply it to OMI spectra and perform an initial validation using independent information on the ozone profile from other sources (sonde, lidar).
Improving Methane Emission estimates Aided by Satellite data (IMEAS) (EO-087, new project)
PI: Dr M. Krol, IMAU
This project aims at a better quantification of the global methane budget. The main focus is on the inter-annual variability of the methane emissions from wetlands and biomass burning. Understanding of the methane budget is important because methane is an important greenhouse gas and its global atmospheric burden seems to stabilise since the last few years. A key question is whether this stabilisation is caused by a reduction in the methane sources, or by an increase in the methane sink (mainly oxidation by OH). A better understanding of the methane budget is needed for predictions of the role of methane in global warming. The detection of methane from satellite platforms would help to constrain the methane budget further. An improved model estimate of methane columns is needed to validate and improve existing retrieval algorithms based on SCIAMACHY data. The objectives of the current proposal are:
• To couple comprehensive biosphere models of methane production to a chemistry transport model
• To use satellite measurements to obtain information about important methane sources, like biomass burning and wetland distributions
• To compare methane columns, simulated with the improved and optimised emission estimates, to SCIAMACHY methane columns
• To detect possible systematic differences between the modelled and measured methane columns with the aim to improve the methane retrieval algorithm
Assessment of the global distribution of the tropospheric OH-radicals from GOME-observations (EO-013)
PI: Prof Dr P. Builtjes, IMAU
This project was successfully completed in 2002. The main objective of the investigation was the evaluation of the oxidising capacity of the troposphere by means of the indirect determination of the tropospheric OH production and budget by using GOME data and model calculations.
In the begining of the project the GOME spectra were reviewed and simulated with a radiative transfer model, and a (software) infrastructure was created to read and process GOME data. The impact of ozone layer depletion was studied. An ozone depletion climatology was created and the changes in photodossociation rates and tropospheric composition were investigated. The results were published in Atmospheric Environment in 2001.
A comparison was done on the four currently available cloud retrieval methods: OCRA (Kuze and Chance, JGR, 1994), PCRA (Kurosu et al, ESAMS'99 Vol 2, 1999), OCRA (Loyola, IGARSS'98 Digest Vol II, 1998) and FRESCO (Koelemeijer and Stammes, JGR, 2001). It was concluded that all four methods retrieve clouds with a high optical thickness. The FRESCO method was found to be currently the best cloud retrieval method for GOME data because of the joint retrieval of the cloud fraction and the cloud top height. The results from this study are publicly available online at the Atmospherical Chemistry and Physics (discussions) website. A study was performed to estimate the errors caused by the 'effective cloud fraction' from the cloud retrieval methods to the primary production of OH radicals. It was found that the local primary production of OH in a 'thick effective cloud' situation can differ considerably to a 'thin effective cloud' with the same radiation at the Top of the Atmosphere. These differences increase with a higher cloud top height and a larger Solar Zenith Angle. This study is accepted for publication in JGR - atmospheres. A methodology was demonstrated to calculate the primary OH production in the troposphere based on GOME satellite data. This methodology was improved, to use FRESCO cloud data and SSP water vapour column data, and then used to calculate the primary OH production for three consecutive days, providing a global coverage. The calculated data were analysed and a quantative uncertainty was attached to the results. This work is in final stages of preparation and will be submitted to a journal shortly.
The calculated tropospheric primary OH production was compared to output of the TM3 Chemical Transport Model, and the agreements/differences were analysed. The calculated OH production values using GOME retrieved data often agrees within the error margin to the TM3 results. This work is in stage of preparation and will be submitted to a journal when ready.
SCIAMACHY Interpretation and Validation Support (EO-019)
PI: Prof Dr H.M. Kelder, KNMI
This project was successfully completed in 2001. The aims of the project were to provide scientific support for preparing the validation and interpretation of SCIAMACHY data; to perform preparatory studies on retrieval algorithms and other measurements for the validation of SCIAMACHY data and finally to perform validation and interpretation of SCIAMACHY data. The launch of ENVISAT has been postponed to spring 2002. Therefore the validation and interpretation of actual SCIAMACHY data was not possible during the course of the project. The emphasis of the project has been on the preparation of the validation, development of trace gas retrieval algorithms and validation tools.
The SCIAMACHY Validation and Interpretation Group (SCIAVALIG) was set up. Various documents were completed, the most recent being the SCIAMACHY detailed validation plan. It will be published just before launch and it explains in detail what will be done during the commissioning and main validation phase. In addition the long-term validation is planned in more detail. A living web site for all SCIAMACHY validation teams has been set up (http://www.knmi.nl/sciamachy-validation/) including a discussion-board for validation results, time schedules, documentation, links and other information.
The validation of GOME ozone observations with a dynamic data assimilation model, which was started in a previous SRON project, was finalized. The data assimilation technique enabled the finding of inconsistencies in the ozone data depending on viewing direction of the GOME instrument. The development of retrieval algorithms for ozone columns from raw GOME data (level 0-1 and level 1-2) was the main area of research.
Sensitivity studies are performed on the effect of algorithm changes to the GOME total ozone columns and to the ozone profiles. Based on these studies level 0 - 1 algorithms have been developed for a set of small spectral windows. A study of the behaviour of the peltier cooler interference has led to the development of a new correction algorithm for this noise term. New algorithms for radiance-degradation correction, PMD-degradation, polarisation correction and wavelength calibration have been implemented.
Also level 1-2 algorithms have been developed based on the above sensitivity studies. The development of these algorithms resulted in knowledge about all steps in the GOME level 0-1 and level 1-2 processing, which are comparable to the SCIAMACHY processing. This knowledge will make it easier to find possible causes for errors in the SCIAMACHY data found during validation.
An atmospheric radiative transfer model developed at VU and KNMI (GAP) is transferred from a plane-parallel model to a semi-spherical model, to take properly into account the change in geometry for high solar zenith angles. This model is used to calculate air-mass factors for DOAS retrieval. A new method for taking into account the different viewing angles during the integration time is developed, which gives more accurate effective air-mass factors than are used in the GOME Data Processor (GDP).
It is shown that the DOAS algorithm is very sensitive to the assumed effective temperature. An effective temperature climatology based on ECMWF data and on an ozone climatology (made of ozone sonde data) is derived. Using this climatology instead of the one used in the GDP is shown to give rise to ozone column differences between 0 and 10 percent (depending on latitude).
SCIAMACHY Science Plan - Support SCIAMACHY Co-PI (EO-020)
PI: Dr A.P.H. Goede, SRON/KNMI
The first scientific results coming out of SCIAMACHY data have been published this year at several international conferences and reviewed articles. These results have caused a major impact on environmental research data. For example, the KNMI NO2 global troposphere column picture hit the headlines of several European daily news papers and TV broadcast in October 2004.
These results have allowed interpretation of the SCIAMACHY data by means of models and application of these data in operational GMES Services*). Assessment of the quality of the data has also provided valuable input for the definition of a follow-on Atmospheric Composition satellite mission***).
Interpretation has taken place in the larger context of the European Commission RTD project EVERGREEN**) aimed at retrieving greenhouse gas emission data by means of inverse modelling. The inverse modelling technique is based on a four dimensional variational data assimilation technique allowing optimisation of both retrieved concentration distributions and emissions. Sensitivity studies have been carried out to quantify the influence of SCIAMACHY data accuracy, the presence of clouds and to optimise the integration time period of SCIAMACHY observations. Early comparison of model results with SCIAMACHY observations show good agreement except for some significant deviations in methane source strength in the tropics occurring in the August to November 2003 integrated time period. This discrepancy is thought to be real and is attributed to hitherto unaccounted for emissions in the tropical areas where the ground-based observation network is only sparse.
*) PROMOTE, ESA GMES Service Element for Atmosphere. GMES (Global Monitoring for Environment and Security) is a joint EC/ESA programme
**) EVERGREEN, a 5th Framework :Programme Research and Technology Development project of the European Commission
***) CAPACITY, ESA General Studies project for the definition of an Operational Atmospheric Composition Monitoring satellite mission based on GMES service requirements
Satellite observations for development of global chemistry transport
model (EO-022)
PI: Dr A.P.H. Goede, SRON/KNMI
In order to interpret satellite remote sensing data, a three-dimensional chemistry transport model TM3-stratosphere with global coverage has been developed in collaboration with the Institute for Marine and Atmospheric Research Utrecht (IMAU) and KNMI.
In preparation for the upcoming ENVISAT launch the 3-d chemistry transport model TM3 was changed, modified and improved to enable a better analysis when comparing SCIAMACHY-data and model results. The TM3 model had already been tested and validated against satellite (GOME) and remote sensing (ASUR and SFINX) data as part of this study (and vice versa). The changes in TM3 are based on these and other results. The new model, called TM5, was developed in close cooperation with, and partially at Utrecht University (IMAU), and is significantly different from the "old" TM3 model. TM5 has the possibility of zooming with a high spatial horizontal resolution of 1°x1° for a limited region. It has been vertically extended from 10 hPa (30 km) to 0.1 hPa (80 km) and the number of vertical levels for the tropopause region and lower stratosphere has been increased, thus obtaining a higher vertical resolution in this region. This enables more detailed studies for this region. Furthermore, a new chemical solver for the gas phase chemistry was implemented. Implemented are also a better description of ozone chemistry for the middle and higher stratosphere (> 30 km altitude), and of tropospheric hydrocarbon chemistry (< 5 km altitude), which is important for carbon monoxide, carbon dioxide and methane. Emissions of atmospheric trace gases have been adjusted and updated according to the EDGAR database. A new parameterisation for NOx emissions from lightning has been implemented. Also, the parameterisations for the surface removal of trace gases and the removal by rain and clouds have been improved to better represent these processes in the model. A new scheme for the photodissociation of trace gases was implemented.
Currently the model is up and running with all aforementioned changes, updates and adjustments, as of the end of November 2001. Testing TM5, especially the stratospheric part, will be subject of future studies (using SCIAMACHY, GOME and possibly ASUR data). The parallelization and the implementation of aerosols will be done at IMAU.
Emission and dissociation of molecular nitrogen in the earth's atmosphere (EO-023)
PI: Prof Dr W.J. van der Zande, FOM-AMOLF/KUN
The middle and upper atmosphere may reveal its secrets by atoms and molecules that spontaneously emit radiation that can be detected with satellite instruments, balloons or on the ground in the case of visible radiation. This emission contains information on the composition, the temperature and the density of important species. In this area of research, our databases are highly incomplete. In the case of nitrogen many electronic states have not been discovered or have not been characterized. States that may be formed upon electron or photon impact on nitrogen gas in the atmosphere. In the case of the red and green oxygen atom airglow, electron collisions on molecular oxygen positive ions are important at elevated altitudes. Using technology developed at the FOM-Institute for Atomic and Molecular Physics and applied at the heavy ion storage ring CRYRING in Stockholm, electron collisions have been measured under circumstances simulating our upper atmosphere, concentrating in the last year at the effects of non-local thermal equilibrium; how does the airglow change if excited molecular ions are present in the atmosphere. The limb measurements from SCIAMACHY may be a tool for routine observations of airglow emissions.
In the coming year, the laboratory studies of molecular oxygen and nitric oxide will be finalized and it will be attempted to check the information content of SCIAMACHY on this aspect.
Cloud absoption retrieval from the near-IR channels of SCIAMACHY (CARIS)(EO-025)
PI: Dr P. Stammes, KNMI
This project was successfully completed in 2001. The objective of this study was to develop, test, and evaluate an algorithm to retrieve the phase of clouds (water/ice) and the size of cloud particles from SCIAMACHY spectral measurements in the near-IR, especially around 1.6 µm. However, since the launch of ENVISAT was postponed, actual SCIAMACHY measurements could not be used for testing and evaluation of the algorithm. Therefore, Along Track Scanning Radiometer-2 (ATSR-2) and Airborne Visible and InfraRed Imaging Spectrometer (AVIRIS) measurements were used for these purposes. Using measurements from these instruments offered the best possible opportunity for the preparation of SCIAMACHY cloud property retrievals.
By finishing this project, we are an important step closer to the global retrieval of macro- and microphysical cloud properties, which is of great relevance to a better understanding of the role of clouds in the climate system. The cloud products that will be derived from SCIAMACHY measurements in the future (i.e. cloud phase, particle size, and optical thickness) will be of great importance for the validation of products obtained from climate and weather models.
The work on cloud-property retrieval from SCIAMACHY will be continued in another GO-project: SCIA-CIRRUS (EO-047).
Sensor synergy study for the Earth Radiation Mission (EO-026)
PI: Dr A.C.A.P. van Lammeren, KNMI
This project was successfully completed and the original research aims have been achieved. The project has clearly demonstrated the importance of combined lidar/radar observations of clouds. This is the only way to reliably measure vertical profiles of cloud parameters (macro- and micro-physical). A summary of the results is given below.
In the first phase of the project, the SSA's for particle size retrieval (from lidar/radar), cloud boundaries (from lidar/radar) and IR emissivity (from lidar/IR radiometer) were developed. A selection of cases from the CLARA campaign was used to test and improve these algorithms.
Also, in co-operation with a related ESA project, some work has been done on the spatial and temporal variability. Results of this study indicate that in the split mission scenario too much information may be lost in order to extract microphysical information on the clouds.
Use of air-borne data (air-borne radar, lidar and in situ measurements from CLARE'98) allowed a better comparison between the results of the particle size retrieval algorithm and in situ particle size measurements. The results show that the retrieval algorithm works well for the cases studied. Addition to the algorithm of a correction for multiple scattering improved the results.
In the case of cloud boundaries, applying the algorithm to CLARE'98 data confirmed the conclusion that in order to determine cloud boundaries, both radar and lidar are necessary. Next, the algorithm to determine the IR emissivity is applied to space-based observation. The space-based lidar (LITE) is used to determine the cloud boundaries. The IR Meteosat channel is used for the IR sky temperature, and NWP reanalysis is used to obtain a temperature profile of the atmosphere. The results show that it is indeed possible to determine the cloud IR emissivity from such observations, although independent validation of the results is difficult.
Regular meetings with ESA representatives have been held to communicate the results. Dr A. van Lammeren was a member of the ERM Advisory Group. Results were used in the ERM assessment report and in the presentation on ERM during "The Four Candidate Earth Explorer Core Missions Consultative Workshop" in Granada, 12 - 14 October 1999. Despite a positive evaluation from the ESA's Earth Science Advisory Committee (ESAC) ERM was not selected as an earth explorer core mission. At this moment the concept of ERM is further developed in a joint ESA-NASDA mission called EarthCARE. The EarthCARE mission has recently been selected for a pre-phase A study.
During the project the PI of the NASA CloudSat satellite has shown interest in the project results. CloudSat will fly the first cloud radar in space and will be launched in 2003. Dr A. van Lammeren is now a member of the CloudSat science team. KNMI has been asked to take the lead in the development of lidar/radar synergy algorithms for CloudSat and PICASSO/CENA (lidar in space).
SCIAMACHY polarisation-correction and validation (EO-027)
PI: Dr P. Stammes, KNMI
The initial aim of this project is to prepare for and to perform the validation of radiance and polarisation data of SCIAMACHY, and to improve the polarisation-correction approach of SCIAMACHY. An important subgoal is to improve the GOME polarisation correction, on which the SCIAMACHY polarisation correction will be based.
The initial aim of this project was to prepare for and to perform the validation of radiance and polarisation data of SCIAMACHY, and to improve the polarisation-correction approach of SCIAMACHY. An important subgoal is to improve the GOME polarisation correction, on which the SCIAMACHY polarisation correction will be based.
An improved UV polarisation correction was implemented and is currently in use for the GOFAP project (GOME fast delivery product, used by KNMI for near real time retrieval of atmospheric parameters, see http://www.knmi.nl/gome_fd/.
A radically new approach to VIS and NIR polarisation correction was developed and shown to be far superior to the operational method for GOME. This new method is also applicable to both GOME-2 and SCIAMACHY.
In the previous report, a new polarisation correction for UV radiances, based on a parametrisation was mentioned. This parametrisation has been expanded (it now also includes viewing direction and scene albedo) and was shown to reduce errors in UV radiances significantly. It has been implemented for the GOME fast delivery product: GOFAP project. The parametrisation will also be used for the GOME-2 polarisation correction. A journal paper addressing this parametrisation will be published shortly.
Work on a radically different polarisation correction that is far superior to the current operational method was also continued. The previously mentioned improved UV polarisation correction is part of this new method. However, at the heart of the new method is not a parametrisation scheme but a novel way to take the spectral behaviour of polarisation into account in a consistent and efficient manner. This method is very flexible and far superior to any other method currently available. This method allows space-borne spectrometers to make full use of their potential (in regard to polarisation). It has been extensively tested, using modelled spectra for all kinds of surface and atmospheric scenarios.
The polarisation correction of SCIAMACHY is more complicated than that of GOME, as more information is needed. Due to the poor wavelength coverage of certain polarisation measurements (U), SCIAMACHY radiances run the risk of being inaccurate. We have developed a new method that uses the spectral shape of Q as a proxy for the spectral shape of U. The results are most promising. This is an entirely new approach to the SCIAMACHY polarisation correction. A technical KNMI report will soon be published.
The polarised radiative transfer code DAK (now DAK 2.4) was extended with routines that allow the use of spectrally varying scattering matrices. This new code was extensively tested and documented. It has been used to calculate realistic earthshine spectra for cloudy and aerosol scenes. These spectra were necessary to test the new polarisation correction method.
Due to the delay of the launch of ENVISAT, the first SCIAMACHY data will only be available after the end of this project. The validation research aim of this project can therefore not be reached. However, we have focussed on developing the polarisation-correction method for both GOME and SCIAMACHY. An extension of the project is needed to perform the validation part. This extension will be realized in another GO-project, EO-054 (SCIA-POLARISATION).
Improvement of GOME calibration keydata (EO-029)
PI: Prof Dr I. Aben, SRON
During the past year the work focused on in-flight monitoring. GOME performs daily solar observations. In the visible wavelength range, and in the UV during low solar activity, the solar output can be assumed constant in time. This means that the sun can be used as a calibration source to monitor instrument performance in time. SRON supported work done at DLR to further improve the degradation correction algorithm based on the use of the sun. Most of the work this year, however, focused on identifying which part of the instrument optical path is affected by degradation and to what extent. The instrument optical path used for solar observations is somewhat different from the moon and earth observations, because during solar observations the light enters through the calibration unit (incl. a diffuser) and is then reflected by the scan mirror into the main spectrometer. During moon and earth observations the light directly enters the spectrometer through the scan mirror, where the incident angle on the scan mirror is different in all cases. Analysis of moon data showed a different degradation behaviour in time compared to the solar observations. This result emphasized the need to monitor earth radiances in time. Although not a trivial matter, we were able to identify a GOME earth radiance data subset for which relatively constant output is expected in time. Specific regions on the earth with rather stable ozone conditions are selected and the radiances 260-295 nm are compared in time.The analysis is limited at the short wavelength end by low signal while at longer wavelengths small variations in ozone would have too much impact on the variation in radiance levels. Comparisons until end of 1998 show almost no difference between earth and solar (ir)radiance degradation. Later comparisons, however, show substantially less degradation of earth radiances compared to solar irradiances thereby severely affecting Earth reflectivity values. GOME Earth reflectivities (@265 nm) at the end of 1999 are ~30% higher compared to GOME 1995-1998 period resulting in huge (±100%) errors in stratospheric ozone profiles retrieved from these data. These results are currently being compared with the outcome of a different analysis by R. van der A (KNMI) that compares GOME data with an ozone climatology. The final quantification of the effect in time is crucial for, a.o., ozone profile retrieval from GOME data.
The different degradation behaviour as observed in GOME solar, moon and earth observations pointed at degradation of the scan mirror. A physical model was developed to model this degradation of the scan mirror performance. In the model the coating on the scan mirror is assumed to be porous. Under ambient on-ground pre-flight conditions the pores are filled with water. The water disappears upon evacuation after launch. The model qualitatively predicts the so-called UV-parabola which was the observation of a rather dramatic wavelength dependent change (~15% @350 nm) in radiance response directly after launch. The performance degradation in time, which occurs on a much longer time scale, is modelled by filling in the pores with some contaminant and successively the formation of an additional contamination layer on top of the coating. This model qualitatively explains the degradation features observed in the different GOME observation modes and thus reveals the main degradation mechanism.
Chemistry and transport of carbon monoxide (CO) in the troposphere (EO-031)
PI: Dr G.J. Roelofs/Dr M. Krol, IMAU
Chemistry transport models (CTMs) are used to increase our understanding of the processes that influence the chemical composition of our atmosphere. In such models the transport, sources and sinks, as well as the reactions amongst the chemically active compounds are described in a computer model. Although many processes are uncertain, the validation of the model with direct measurements in the atmosphere shows that our understanding of the governing processes is quickly increasing. Faster computers enable a higher resolution of the models and the possibility to include more processes. Also the emissions of chemically active species have been better quantified and more information about the chemical interactions (e.g. rate constants) becomes available.
The aim of this project, which was successfully completed in 2002, was to increase our understanding of the budget of carbon monoxide (CO) in the lower parts of the atmosphere. CO plays an important role in the chemistry of the atmosphere. It is emitted by natural and anthropogenic sources. However, the emitted amounts of e.g. the burning of biomass and wild fires is rather uncertain. A large part of the project concentrated on the development of a CTM that is able to zoom in over a specific region (e.g. Europe). The rationale behind this approach is that the chemical lifetime of many species is rather long (months to years) so that the long-range (global) transport of these species should be considered. At the same time, a comparison of modelled concentrations of e.g. CO with ground-based, aircraft, or satellite data requires a higher resolution than can be obtained with a global model. Why simulate the detailed, high resolution chemistry of CO at the south pole, if we are interested in the European scale?
Within the project, such a zoom model has been built and its first results have been compared to surface and aircraft CO measurements. It appears that the model is very well able to simulate the CO fluctuations that are caused by weather changes. However, a comparison to measurements made during the MINOS campaign on Crete in 2001 shows that the emissions of CO are probably the most uncertain part in its budget. Large amounts of CO are emitted by processes like agricultural waste burning and forest fires, but the exact quantities are unknown. It is therefore recommended (and first efforts in that direction have been made) to better quantify (CO) emissions by using observations. Satellite observations of CO columns can certainly play an important role. Our model results show that the local variations of the total column of CO due to emissions and transport processes are in the order of 20-30%. If satellites are indeed able to measure the total columns of CO with an accuracy of 10%, the wealth of data that will be provided by such an observation platform will be of great value in the determination of CO sources in e.g. Europe.
Retrieval algorithms of micro-physical cloud parameters for the Earth Radiation Mission (EO-035)
PI: Dr Ir H.W.J. Russchenberg, DUT-IRCTR
Knowledge of water clouds is very important for climate studies, but retrieval of their micro-physical properties from ground-based and space-based remote sensors is difficult and challenges the technological possibilities of today. This is partly due to the limitations of instruments and partly due to physical processes in water clouds that obscure direct relationships between remote sensing observables and the cloud parameters. More specifically, the 'large droplet issue' in water clouds - just a few small drizzle droplets in water clouds may cause a radar signal that is much stronger than that of the radiatively important fraction of the cloud - makes the direct retrieval of microphysical properties with one sensor practically impossible.
In this project a new technique was developed that overcomes these limitations by means of sensor synergy: radar in combination with lidar can be used to classify clouds in types that contain drizzle droplets and those that do not. Appropriate retrieval algorithms can then be applied to each category. This technique developed on the base of the study of in-situ droplet size spectra in water clouds that were measured with aircraft-mounted probes during a few field campaigns in the different geographical regions around the globe and in the different types of clouds. The technique tested with ground-based simultaneously measured radar and lidar data and the retrieval results validated though the comparison with the liquid water path, which was independently measured with microwave radiometer. The different methods of the spatial and temporal averaging of the real data were use to simulate radar and lidar observations from space.
The developed method can be used for the water cloud microphysics retrieval during the successor of the Earth Radiation Mission (ERM), called Earth CARE, which is currently being investigated by ESA in co-operation with NASDA, with a payload comprising a lidar, radar, imager and broadband radiometer. In nearest future it can be applied to the data from the CloudSat and CALIPSO joint satellite mission.
Laboratory measurements and spectroscopic modelling in support of retrieval of Greenhouse gases from the SCIAMACHY infrared channels (EO-036)
PI: Prof Dr W. Ubachs, VU, Co-PI: Dr J. Schrijver, SRON
In 2003 further measurements were performed to determine extinction cross sections of atmospheric molecules to cover the entire visible wavelength range from the blue to red. These extinction cross sections for N2, Ar, SF6 and CO2 molecules relate entirely to the Rayleigh scattering cross sections, but it is for the first time that values are determined in a direct way. The obtained results can be compared with data calculated from refractive indices, measured previously at high accuracy. However the Rayleigh cross section also depends on the so-called King factor, which is a measure for the anisotropy of the molecule. Particularly for CO2 the King factor is large and the present measurements allow for a determination of the King factor.
Secondly the collision-induced phenomena for O2-O2 pairs were further investigated. Room temperature measurements for the “blue phenomenon” at 477 nm were completed. A new cell was constructed for the below room temperature regime and measurements on the “yellow” and “red” features at 577 and 630 nm were studied in a temperature regime down to –100 oC. The results indicate a pressure dependence for these collisional absorption features. Currently measurements on the temperature dependence of the 477 nm collision-induced absorption phenomenon are being performed; these data are of relevance since this feature is implemented in the OMI retrieval algorithms for determination of cloud top heights.
Alongside a new laser system, developed at the Laser Centre VU, for the generation of ultra-narrowband laser pulses in the deep-UV (below and around 200 nm) has been implemented in cavity-ring down studies to re-examine the Schuman-Runge absorption features. In particular the weak (0,0) band has been recorded at the highest resolution ever. This development sets the stage for a new research programme to investigate atmospheric molecules with the highest sensitivity and resolution in the deep ultraviolet.
Development, validation and application of aerosol retrieval algorithms using GOME, ATSR, SCIAMACHY, and AATSR (EO-037)
PI: Prof Dr G. de Leeuw, TNO-FEL
Atmospheric aerosols are of great importance because of their role in long-range transport of pollutants, their impacts on human health, the Earth’s climate, visibility and photosynthetic active radiation (PAR). Recent estimates show an uncertainty of 40-200% in the direct radiative effect of aerosols, whereas the indirect effect, through their influence on cloud radiative properties, cannot be quantified with any certainty. In Europe, health effects are currently based on PM10, i.e. the mass concentration of dry aerosol for particles smaller than 10 μm in diameter; other standards under discussion are PM2.5 and PM1. Estimates of the effects of particulate matter in ambient air in The Netherlands amount to 1700 premature deaths per year, up to 10,000 to 15,000 per year due to long-term exposure. The occurrence of dust in the atmosphere renders people more susceptible to infections. Terrestrial and maritime ecosystems are affected through the deposition of pollutants and nutrients. Aerosols influence atmospheric chemistry through heterogeneous chemistry and photolysis rates, which affects the stratospheric ozone layer. For these and other reasons dedicated monitoring of aerosol concentrations and properties at the European and global scales is required.
The work undertaken in this project was aimed at the use of satellites to determine the aerosol composition and their effects on climate and air quality. The detection of aerosols by satellites is based on their effect on the reflected solar radiation observed at the top of the atmosphere. Contributions from the reflections by the Earth’s surface, atmospheric gases and aerosol can be separated by using available information on the surface properties and the effects of gases. Thus the effect of aerosols remains and can then be used to retrieve aerosol properties. Algorithms to accomplish this have been developed and successfully applied to satellite observations over Europe with the ATSR-2 instrument. From comparison with model results information on aerosol chemical composition, PM2.5 and climate effects could be obtained. The presence of absorbing aerosol such as black carbon and desert dust, renders the retrieval more difficult. This problem has been approached over areas for which extensive data sets on such aerosols are available, i.e. the Indian continent and adjacent oceans and the southern part of Africa. The resulting algorithms have been applied to derive maps of the mean AOD, the wavelength dependence of the AOD, the aerosol mixture, and the contribution of black carbon to the total aerosol mass over South Asia and the Indian Ocean. Similar principles have been used to develop an algorithm for aerosol retrieval over sea and over land from SCIAMACHY data. To test this algorithm it has been applied to GOME data to provide a global map of AOD over land and over water. This project was successfully completed in 2003.
Occurrence and modelling of the 1.7 year oscillation in stratospheric ozone (EO-038)
PI: Dr F.J.M. Alkemade, IMAU
The isolation of the signal of the 1.7 year oscillation, responsible for about 25% of the (deseasonalized) total variability in a quarter century of successive TOMS (total ozone monitoring spectrometer) data has allowed for a very sophisticated MLR (multi linear regression) analysis of total ozone. This leaves a so-called residue (i.e. the unexplained variability that remains in the data after the data are corrected for a number of natural cycles and other known influences) that contains less noise and thus more meaningful information about possible physical causes of the variability in this residue.
During the last year the research was primarily aimed at applying this analysis to the interconnection between stratospheric ozone, changing temperature profiles (due to the increase of well-mixed greenhouse gases as well as to the depletion of ozone) and changes in tropopause height.
Ultimately this may lead to more reliable techniques for diagnosing present and future ozone changes (e.g. to distinguish between the effects of greenhouse-related temperature changes and the effects of CFC-related ozone depletion). Also it will possibly allow for an earlier and more reliable detection of the so-called 'turn-around' point, the moment at which the ozone-layer will start to recover.
Studies of Assimilated SCIAMACHY data (SASCIA) (EO-040)
PI: Dr P.F.J. van Velthoven, KNMI
To obtain insight in the state and behaviour of the atmosphere, scientists have two tools available: models and measurements. The SASCIA project aimed to combine model simulations and satellite measurements of ozone in order to obtain the best available insight in the concentrations of this important constituent over long periods. The quality of the satellite measurements has been improved over the years, such that now for the first time high quality measurements at different altitudes have become available on a regular base.
In combination with model simulations, these measurements have given insight in how ozone is transported through the atmosphere.
Development of a radiative transfer model for SCIAMACHY limb measurements (SCIARALI) - Phase I (EO-044)
PI: Prof Dr H. Kelder, KNMI
This project was successfully completed in 2001. The aim of the project was the development of a radiative transfer model (RTM) that calculates radiances at limb view and derivatives with respect to atmospheric properties (Jacobian or weighting functions) at the top of the earth's atmosphere (TOA). This model is to be applied in the retrieval of vertical profiles for a number of important gaseous constituents of the Earth's atmosphere from limb measurements of the SCIAMACHY satellite-based instrument.
During the one year period of the project, the main attention has been on the development of an RTM in spherical geometry for single scattered radiation (suitable for both limb and nadir view) as well as the study of methods to be applied for the successive inclusion of multiple scattering processes in the RTM. The actual inclusion of multiple scattering processes in the RTM will be part of the follow-up research project EO-051.
A joint SRON/KNMI atmospheric model was set up in order to calculate the optical properties of a model atmosphere based on different AFGL atmospheric profiles. Ozone absorption, Rayleigh scattering as well as aerosol absorption and scattering properties can be included. This common set-up was chosen in order to ease model comparison in the second phase of SCIARALI.
At KNMI, a radiative transfer code for single scattered solar radiation in a spherical geometry of the Earth's atmosphere was completed. The code is suitable for arbitrary observation angles and therewith conforms applicability for both limb and nadir observations. The RTM produces radiances leaving the top of the atmosphere. In concern with the aforementioned atmospheric model, this single scattering radiative transfer model provides a reliable set-up for the simulation of the near-infrared SCIAMACHY measurements for clear sky atmospheres, where multiple-scattering processes can be neglected.
The module to calculate the Jacobian of the radiance with respect to atmospheric properties (weighting functions) is close to completion. Preparatory interfaces for the inclusion of multiple scattering processes in the RTM have been created.
For the second phase of SCIARALI, the inclusion of multiple scattering processes in the RTM, two approaches have been investigated: the LIDORT model and a pseudo-spherical extension of the existing radiative transfer model LIRA.
LITE for ADM (EO-045)
PI: Dr A. Stoffelen, KNMI
In this project, which was successfully completed in 2002, LITE data (Lidar In-space Technology Experiment) was used to shed more light on the Atmospheric Dynamics Mission (ADM). The ESA ADM Doppler Wind Lidar (DWL) operates at 355 nm and detects both molecular, and cloud and aerosol scattering. Distribution of cloud and aerosol have been recognised as crucial DWL parameters. In the processing of the data, to obtain wind component vertical profiles, effects of cloud dynamics and spatial distributions of cloud and aerosol may cause systematic errors. The goal of this project was to shed more light on the spatial representativeness and sampling errors and to quantify them using LITE data.
The most time-consuming part of the project involved the development of a novel method to retrieve aerosol backscatter characteristics from LITE data. It was found that existing retrieval methods from literature were not useful. This requires a robust retrieval method that is capable to handle noisy data. A novel method was developed using data of the three wavelengths simultaneously. Moreover, statistics for the aerosol backscatter-to-attenuation ratios were derived at all three wavelengths. Retrieved aerosol backscatter statistics at low resolution were verified against literature and found in close agreement.
The LITE retrieved aerosol backscatter characteristics were used in the second part of the project to assess potential systematic errors in spaceborne lidar measured winds in heterogeneous atmospheres. From the results it was concluded that aerosol backscatter variability along a 50 km satellite ground track is generally small in cloud-free conditions, implying that lidar measured wind profiles are generally very well representative for the 50 km along track sampling. Even in cloudy conditions the horizontal representativeness is generally good. Only in cases where part of the 50 km sampling interval is covered with clouds, positioning errors of the wind profile may be significant. However, since ADM oversamples the 50 km horizontal integration length by 14 accumulations, post-processing schemes may discriminate between cloudy and cloud-free returns.
The ADM-Aeolus vertical sampling of 1 km appears more problematic. Aerosol variability over 1 km in the vertical is generally significant even in cloud-free conditions. For the Mie (aerosol) channel height assignment errors of about 75 m in the lower stratosphere were found and about 50 m in the mid and upper troposphere and lower stratosphere. For the typical windshears of 4 m/s per 1000 m this implies systematic errors of about 0.2 to 0.3 m/s in lidar measured winds. Errors from the Rayleigh (molecular) channel, due to aerosol extinction, are much smaller. For cloudy scenes the results are worse with errors up to 125 m (0.5 m/s) in the lower atmosphere and about 75 m (0.3 m/s) in the upper atmosphere for the Mie channel. Again, the results are less dramatic for the Rayleigh channel.
Next, we collocated LITE atmospheric conditions with ECMWF meteorological model fields in a couple of case studies to simulate errors of lidar measured winds in realistic atmospheric conditions. Areas of large variability of aerosol distribution over short vertical distances were found that may give large errors in retrieved winds in cases of large wind gradients. An important conclusion form the results is that oversampling in the vertical would be very beneficial to reduce systematic errors. Moreover, advanced ground-processing schemes should be able to discriminate between cloudy and cloud-free accumulations and both should be treated separately for wind retrieval.
Data assimilation of radiances for ozone profile extimation (DROP) (EO-048)
PI: Prof Dr H.M. Kelder, KNMI
In the DROP project, which was completed in 2003, scientists in the Atmospheric Composition department of KNMI are developing a new method to assess the 3D distribution of ozone in the atmosphere. This method directly combines information from an atmospheric chemistry model with satellite measurements of sunlight reflected off the earth’s atmosphere. Methods using both model and measured data to improve our knowledge of a system are known as “data assimilation”. This project is the first initiative to directly assimilate UV spectra measured by satellite instruments like GOME, SCIAMACHY or OMI into a chemical transport model, thereby making optimal use of the profile (vertical) information contained in the measurement. The usual practice in this field is to first determine (retrieve) a total ozone column or an ozone profile from the radiance data, and assimilate this estimate into a model. The results of this retrieval depend strongly on initial assumptions made about the ozone or temperature profile. Radiance assimilation is a big advance over these methods, as it does not need these external assumptions about state of the atmosphere. Hence, with data assimilation of radiances into a state-of-the-art ozone model we expect to be able to produce a very high quality data set of ozone.
Major progress was made in the development of the components of the assimilation algorithm; what remains to be done is the integration of these into a working system that is fast enough for near-real-time data processing. For this to succeed more research into details of the analysis system is necessary.
Development of a radiative transfer model for SCIAMACHY limb measurements - Phase II (SCIARALI - Phase II) (EO-051)
PI: Prof Dr H.M. Kelder, KNMI
This project was successfully completed in 2003.
The SCIAMACHY instrument aboard the Envisat satellite has the unprecedented capability to observe the Earth’s atmosphere in two different ways: looking straight down (nadir mode) and observing the horizon (limb mode). The aim of these observations is to derive information of the distribution of atmospheric chemical constituents, such as ozone.
In a joint effort, KNMI and SRON have developed radiative transfer models and retrieval algorithms for the retrieval of the vertical distribution of trace gases from SCIAMACHY limb observations. For the development of the radiative transfer modelling, it appeared to be necessary to simulate the propagation of light through a spherically shaped model atmosphere. Retrieval involves the comparison of model output with the satellite measurements; the true composition of the atmosphere can be derived in an iterative manner.
To account for curvature of the Earth's atmosphere, two different approaches have been followed. One model aims at a high accuracy by realistically taking the sphericity of the atmosphere into account. The second model seeks a trade-off between accuracy and computational speed by compromising on the way the curvature of the atmosphere is treated: for solar light that is scattered only once by the atmosphere, the sphericity is fully accounted for, whereas for multiply scattered light it is assumed that locally the atmosphere extends horizontally without curvature. At present, the latter of the two models has been incorporated in a retrieval scheme. In the near future, this algorithm will be applied to derive atmospheric trace-gas distributions from actual SCIAMACHY observations.
More information: pb@sron.nl