The task of KNMI within the CLOUDMAP project is to evaluate the potential end-use of CLOUDMAP products and techniques by the European meteorological and climate research community. For this purpose KNMI has performed a "roadshow" among relevant user-institutes in Europe. The roadshow consisted of five meetings, held in five different countries:

• Deutsche Wetterdienst (DWD) at Offenbach, Germany on September 19, 2000.
• Laboratoire de Météorologie Dynamique (LMD) in Paris, France on October 6, 2000.
• The Met Office in Bracknell, UK on October 13, 2000.
• Swedish Meteorological and Hydrological Institute (SMHI) in Norrköping, Sweden on October 19, 2000.
• Royal Netherlands Meteorological Institute (KNMI) in De Bilt, The Netherlands on November 22, 2000.

During these meetings the CLOUDMAP evaluators from KNMI have met potential end-users and experts on cloud information from satellites with the purpose to obtain feedback on the CLOUDMAP project itself and its applicability in real practice. Over a hundred people took part in the five different meetings. The agendas of the one-day meetings were conducted in consultation with the local hosts. The morning part of all meetings was programmed in the same way. First, Hans Roozekrans of KNMI gave a one-hour presentation of the CLOUDMAP project to a larger audience (varying from 10 to 25 people). The CLOUDMAP techniques (stereo matching, O2 A band retrievals, contrail detection), examples of products and some validation results were presented. After coffee break a discussion-round of another hour was held using ten statements, prepared by KNMI. At DWD, LMD and SMHI small meetings with three types of end-users (forecasters, NWP modellers, and climate researchers) were held in the afternoon. At the Met Office it was decided for logistic reasons to deviate from this proven concept and to repeat the morning programme in the afternoon. The morning programme was focused on the use of CLOUDMAP in operational forecasting and the afternoon programme on the use in NWP and climate research. At KNMI no afternoon activities were programmed.

The consulted user communities are asked for their opinions on the success of the CLOUDMAP project. The users were also asked for their opinions on the potential application of CLOUDMAP products in the three target application fields. In this paragraph it is tried to reflect the highest common factor of all these user opinions.

The CLOUDMAP project can not (yet) be considered as a major forward step in satellite cloud meteorology and climatology. However,

• the results of CLOUDMAP are interesting and promising,
• the independent character of the CLOUDMAP products is the key factor for a potential successful application,
• the accuracy of the products is not yet adequately demonstrated.

• Cloud top height (CTH) is not a key parameter in this application field (cloud base height and the temporal development of clouds, e.g. CB's, has higher priority). The future operational meteorological satellites (MSG, Metop) are expected to provide sufficient (both qualitative and quantitative) CTH products for real-time use by forecasters.

• Forecasters show a large interest in microphysical cloud parameters (like droplet phase, radius and concentration) as can be derived using the O2 A band technique (besides other technologies that could do it equally well or perhaps better). The O2 A band technology should be considered for inclusion in future operational meteorological imaging satellite sensors.

• Information on contrails is not relevant for operational forecasting.

• In the near future there will be a need for CTH data, independently observed from satellites, for application in NWP. The availability of a solid error structure of the CTH product is a necessity for successful application in NWP.
• The NWP models are not yet ready for the assimilation of microphysical cloud parameters. First, new parameterisation and assimilation schemes in the NWP models need to be researched and developed. CLOUDMAP products can contribute to this R&D work.
• Information on contrails is not relevant for NWP.

• The climate research community has large interest in independently observed CTH data. However, the community sets high demands to a solid validation of the product before application in climate research will be accepted. At this stage of the CLOUDMAP project the validation of CTH products does not match this requirement.

• Climate researchers show a large interest in microphysical cloud parameters (like droplet phase, radius and concentration) as can be derived using the O2 A band technique. Therefore, together with the operational meteorological community the climate research community strongly supports the idea to consider the inclusion of the O2 A band in future imaging sensors on board operational meteorological satellites.

• There is an ambiguous opinion within the consulted climate research community about the importance of contrail information in climate research and, therefore, also about the applicability of the CLOUDMAP contrail products in climate research.

• The CLOUDMAP products have large potential to be used for the validation/verification of Nowcasting and Climate SAF cloud products. The independent character of the CLOUDMAP products favours this application. However, the problem of collocation (in time and space) of the products will limit this application in practice.
• The CLOUDMAP CTH products form a potential independent data source for height assignment of cloud motion vectors (as used in NWP models).
• The consulted climate researchers stress the need for validation of CLOUDMAP products of tropical climates.
• The CLOUDMAP (1 and 2) results and plans must be soon communicated to the current and future discussion forums within ESA and EUMETSAT that already now (or quite soon) are discussing the payload of next generation operational meteorological missions.
• The climate research community advises (and offers opportunities) to link the CLOUDMAP2 work to the international cloud observation and measuring campaigns that are planned in the coming three years. The benefits will be twofold: opportunities to validate CLOUDMAP products using high quality cloud measurements and CLOUDMAP publicity in the climate research community.

This first meeting of the CLOUDMAP roadshow was a sort of a try-out to experience the effectiveness of the chosen format. The CLOUDMAP representatives from KNMI and the end-users and experts from DWD and EUMETSAT were quite happy with the formula. For both sides the meeting turned out to be very beneficial. Knowledge on new retrieval techniques for cloud information was brought to the end-users and useful feedback on the applicability of CLOUDMAP was given back in return. The sessions with individuals and smaller groups in the afternoon were beneficial too, especially for the settlement of future co-operation on data-collection in the framework of CLOUDMAP2 and the Climate SAF project (of which DWD is host).

The main conclusions of the DWD meeting are:

• The potential of application of CLOUDMAP products in operational forecasting and in NWP is limited at the moment. The forecasters are quite satisfied with the current available satellite products and they have a problem with absorbing more information. At DWD there is hardly any experience with the use of cloud data in relation to NWP modelling.
• The climate researchers and Climate SAF group are more positive on the potential application of CLOUDMAP products. Especially, the link between CLOUDMAP and the SAF work was stressed. The Climate SAF group sees CLOUDMAP as a potential data source for validation of the SAF products.
• The DWD judgement of the CLOUDMAP project: CLOUDMAP is not (yet) a major step forward in satellite remote sensing of clouds.
• EUMETSAT follows the CLOUDMAP work with interest, but is not yet convinced of the applicability of CLOUDMAP techniques to future EUMETSAT missions.

The following persons attended the morning programme of the roadshow meeting at DWD. The initials (in alphabetic order) in the first column are used in this discussion report to identify persons who took part in the discussion. The comments made by the individual participants are not verbatim represented in this report and should not be treated as official.

• JK: the contrail occurrence map raises doubts about the earth surface dependence of the detection efficiency in the contrail detection algorithm. There appears to be a difference in the density of contrails above land and sea.

• DM: Why not comparing the derived cloud top heights (CTH) to other observations, e.g. radiosondes. (This caused the remark that the detection efficiency of CTH in radiosondes depends on the cloud type: water clouds are better detected compared to thin ice clouds).

• WB remarks that the DWD observation site at Lindenberg is willing to support validation of cloud products. At Lindenberg several ground-based instruments continuously observe and measure the clouds.

• WB asks about the satellite instrumentation that is intended to be used in CLOUDMAP2. Reply: among the ones used in CLOUDMAP1: MSG, MISR, MERIS, MODIS, etc.

• The question was raised how to validate the cloud optical thickness products.

Statement 2: The present and future operational satellite cloud data (METEOSAT, MSG, NOAA, Metop) are/will be sufficient for end-users.

Statement 3: Considering the quality is acceptable, end-use of CLOUDMAP products can only be successful if priority is set to the improvement of

Statement 4: CLOUDMAP contrail products are relevant for / applicable in:

Statement 5: Application fields of CLOUDMAP products with the highest probability of value adding are:

Statement 6: In case the requirements to timeliness and data frequency can be met by CLOUDMAP, potential application fields are:

Statement 7: CLOUDMAP products have potential to be used for validation/ calibration of operational satellite cloud products; e.g. MPEF or SAF (NWC, Climate) products.

Statement 8: CLOUDMAP techniques have potential to be applied to or implemented on future meteorological satellites.

Statement 9: At the end CLOUDMAP will contribute to the availability of more accurate cloud information from satellites.

In this part of the programme individual end-users within the three target groups (forecasting, NWP and climate research) were (on-line) introduced to the CLOUDMAP Web site. During this introduction the applicability and limitations of the CLOUDMAP products and techniques were discussed. At the end of each session the end-users were asked to reply to the end-user evaluation questionnaire. Three sessions of an hour each were held.

The forecasters had a positive reaction on the stereo and O2 band CTH products mainly because of its independent character. They had some worries about the occasional difference between the optical and thermal stereo CTH results. For an operational forecaster, having not too much time during the shifts, this will need too much interpretation. Therefore, a more integrated end product is advised.

Applications of CTH in operational forecasting/nowcasting:

• aviation meteorology,
• nowcasting of CB development,
• estimation of fog thickness.

The droplet concentration product is found to be very useful for the objective analysis of NWP models.

Participant: Bodo Ritter (NWP modeller)

The independently derived CTH is found to be useful for calibration and validation of NWP models. Currently, CTH is not directly assimilated in the NWP model of DWD (the EUMETSAT Cloud Motion winds are used). The problem with CTH assimilation is the definition of the CTH. This is very much related to the spectral response of the clouds. In the DWD NWP model three rather broad spectral bands are used to define CTH. If e.g. ATSR stereo CTH data are used for NWP-assimilation a translation of the narrow ATSR bands to the broad bands (as now used in NWP) must be performed.

Participants: Annegret Gratzki, Gabi Wollenweber and Heinrich Woick (the DWD team working on the Climate Monitoring SAF)

The future of the application of space technology in meteorology and climate research was discussed. Heinrich Woick has heard signs within the meteorological community that the future will be dominated by multi-spectral sensors like MODIS and MERIS. This fits well to the CLOUDMAP2 project!

Furthermore, the potential use of CLOUDMAP products in the Climate SAF was discussed. The time path of CLOUDMAP2 fits very well to the time path of the development phase of the Climate SAF. Both projects will end by the end of 2003. Co-operation between both projects seems very logical and beneficial for both. The Climate SAF group identifies itself as an obvious potential end-user of CLOUDMAP.

In this second meeting of the CLOUDMAP roadshow the formula of the first meeting at DWD was used again. Again it worked out quite beneficial for both sides. The introduction to CLOUDMAP in the morning was followed by a fruitful and lively discussion focused on the application of CLOUDMAP in climate research. The sessions with individuals and smaller groups in the afternoon were beneficial too, especially for the settlement of future co-operation on data-collection in the framework of CLOUDMAP2 and the experiments LMD is participating in.

The main conclusions of the LMD meeting are:

• LMD sees potential to apply CLOUDMAP products in its future climate research work. High priority should be set to a proper validation of the products, not only for Europe but also for tropical regions.
• LMD experts doubt the usefulness of contrail products for climate research.

The following persons attended the morning programme of the roadshow meeting at LMD. The initials (in alphabetic order) in the first column are used in this discussion report to identify persons who took part in the discussion. The comments made by the individual participants are not verbatim represented in this report and should not be treated as official.

This report reflects the third trip in the roadshow to the Met Office in Bracknell, UK. The Met Office was chosen because it is one of the major meteorological institutes in Europe and it contains the Hadley Centre, one of the largest climate research centres in Europe. The venue in Bracknell also gave the opportunity to NWP experts of ECMWF to attend the afternoon programme. For the same reason the CLOUDMAP project co-ordinator, Jan-Peter Muller of UCL, was able to attend the meeting, which was very helpful for the success of the meeting. The large interest of people attending the meeting forced a programme deviating from the other four roadshow meetings. This time the meeting was split into two identical programmes, focused on different target groups. An agenda for the one-day meeting was conducted in consultation with the host at the Met Office, Fiona Smith:

Morning programme, focused on operational forecasting:

The main conclusions of the Met Office meeting were:

• There is a need for a proper assessment of the error characteristics of the cloud products, especially for application to NWP.
• Validation of the cloud products, to check the error characteristics are as expected, is considered important for operational applications such as forecasting. However, it was recognised that validation exercises have their limitations.
• In operational forecasting information on the dynamics and cloud patterns is more needed than cloud top information.

The following persons attended the morning programme of the roadshow meeting at the Met Office. The initials (in alphabetic order) in the first column are used in this discussion report to identify persons who took part in the discussion. The comments made by the individual participants are not verbatim represented in this report and should not be treated as official.

Statement 1: End-users require satellite cloud products that are validated, although proper validation of most cloud parameters is difficult or hardly possible.

Statement 2: End-users require satellite cloud products that are independent of additional (e.g. NWP models, synops, etc.) data sources.

Statement 3: The present and future operational satellite cloud data (METEOSAT, MSG, NOAA, Metop) are/will be sufficient for end-users.

Statement 4: Considering the quality is acceptable, operational end-use of CLOUDMAP products can only be successful if priority is set to the improvement of

Statement 5: CLOUDMAP contrail products are relevant for / applicable in:

Statement 6: Application fields of CLOUDMAP products with the highest probability of value adding are:

Statement 7: In case the requirements to timeliness and data frequency can be met by CLOUDMAP products, potential application fields are:

Statement 8: CLOUDMAP products have potential to be used for validation/ calibration of operational satellite cloud products; e.g. MPEF or SAF (Nowcasting, Climate) products.

Statement 9: CLOUDMAP techniques have potential to be applied to or implemented on future meteorological satellites.

Statement 10: At the end CLOUDMAP will contribute to the availability of more accurate and independent cloud information from satellites (the original justification of the CLOUDMAP project).

The following persons attended the afternoon programme of the roadshow meeting at the Met Office. The initials (in alphabetic order) in the first column are used in this discussion report to identify persons who took part in the discussion. The comments made by the individual participants are not verbatim represented in this report and should not be treated as official.

This report reflects the fourth trip in the roadshow to the Swedish Meteorological and Hydrological Institute (SMHI) in Norrköping. SMHI was chosen for several reasons. It is a typical middle size European meteorological centre hosting an operational forecasting department, an excellent NWP modelling group (SMHI is a prominent member of the HIRLAM consortium) and the well-known Rossby Centre (for climate research). Moreover, SMHI is a member of the Nowcasting SAF consortium developing the software for nowcasting products from NOAA and Metop data. Last but not least, SMHI will be partner in the CLOUDMAP2 project and also in the related CLIWANET project.

An agenda for the one-day meeting was conducted in consultation with the host at SMHI, Anke Thoss:

In this fourth and last meeting of the CLOUDMAP roadshow the formula of the first two meetings was recovered and again it worked out quite beneficial for both sides. As at the Met Office the local organiser Anke Thoss had managed to mobilise a quite large audience for the morning session. The afternoon sessions with individuals and smaller groups were beneficial too, especially for the settlement of future co-operation on data-collection in the framework of CLOUDMAP2 and the Nowcasting SAF project (in which SMHI is taking care of the development of NOAA-AVHRR products).

The main conclusions of the SMHI meeting are:

• For use of quantitative cloud products in operational forecasting and in climate research it is very important to provide error bars. For use in NWP the structure of errors is more important (rather then absolute accuracy).
• The independent character of the CLOUDMAP products is very attractive in relation to any application in NWP (assimilation, verification and parameterisation). SMHI has high expectations on the impact of applying satellite cloud data in NWP work and, therefore, will participate in the CLOUDMAP2 project.

The following persons attended the morning programme of the roadshow meeting at SMHI. The initials (in alphabetic order) in the first column are used in this discussion report to identify persons who took part in the discussion. The comments made by the individual participants are not verbatim represented in this report and should not be treated as official.

• A remark was made that relative accuracy of CTH is more needed than absolute accuracy.

• A remark was made that in the case of the ATSR stereo technique users obtain two (or three) CTH results. How to deal with this in real practice?

Statement 1: End-users require satellite cloud products that are validated, although proper validation of most cloud parameters is difficult or hardly possible.

Statement 2: End-users require satellite cloud products that are independent of additional (e.g. NWP models, synops, etc.) data sources.

Statement 3: The present and future operational satellite cloud data (METEOSAT, MSG, NOAA, Metop) are/will be sufficient for end-users.

Statement 4: Considering the quality is acceptable, operational end-use of CLOUDMAP products can only be successful if priority is set to the improvement of

Statement 5: CLOUDMAP contrail products are relevant for / applicable in:

Statement 6: Application fields of CLOUDMAP products with the highest probability of value adding are:

Statement 7: In case the requirements to timeliness and data frequency can be met by CLOUDMAP products, potential application fields are:

Statement 8: CLOUDMAP products have potential to be used for validation/ calibration of operational satellite cloud products; e.g. MPEF or SAF (Nowcasting, Climate) products.

Statement 9: CLOUDMAP techniques have potential to be applied to or implemented on future meteorological satellites.

Statement 10: At the end CLOUDMAP will contribute to the availability of more accurate and independent cloud information from satellites (the original justification of the CLOUDMAP project).

In this part of the programme individual end-users within the three target groups (forecasting, NWP and climate research) were (on-line) introduced to the CLOUDMAP Web site. During this introduction the applicability and limitations of the CLOUDMAP products and techniques were discussed. At the end of each session the end-users were asked to reply to the end-user evaluation questionnaire. Four sessions were held.

Participant: Colin Jones of the Rossby Centre

Colin started to show some work on the cloud parameterisation in a regional (European) climate model. For this purpose a ten-year data set of AVHRR cloud fractions (work of Karl-Göran Karlsson) was used. The limited value of the ISCCP data set was discussed. ISCCP only distinguishes between low and high clouds. For model parameterisation this is much too coarse.

Furthermore, the potential application of CLOUDMAP in climate modelling was discussed.

The vertical resolution of models is smaller than 1 km. When the errors of CLOUDMAP CTH products are in the order of a few 100m than these will be negligible to the model errors, and therefore very interesting for climate modelling.

The CLOUDMAP droplet concentration product looks very interesting. This quantity is important for cloud condensation growth and precipitation.

Cirrus in the deep tropics is very important for climate studies and most climate models under predict these clouds (will the CLOUDMAP2 products have a more global coverage)?
Climate models do need cloud information: clouds and radiation are most important in regional climate models (cloud position, CTH and cloud constituencies). The product should have a spatial resolution better than 50 km and should describe the diurnal cycle.

Finally, the overlap between CLIWANET and CLOUDMAP2 was discussed. Colin Jones emphasises the strong need for co-ordination and co-operation between the two projects. Exchange of data sets will benefit both projects.

The current status of use of satellite data in HIRLAM was discussed. Currently, the following satellite data are used in HIRLAM at SMHI: GPS (TZD), AMSU-A and ATOVS.

Nils asks which stereo matching algorithm will be chosen? This is important to know if stereo CTH will be use in NWP. Impact studies might give an answer to that.

Nils asks whether the O2 A band will be operational available in the near future? NPOES (launch not before 2010) will have an O2 A band. In the mean time MODIS, being an experimental research sensor, will fill the gap.

Nils asks: Does the O2 A band measure integrated water vapour (IWV) above clouds? This is an important parameter for NWP.

Nils asks: What extra information does the O2 A band contain with regard to the WV-channel and AMSU-B? When the underlying surface contains no ice AMSU-B and WV provide similar observations. The O2 A band is complementary to GPS and SSM/I.

Anke foresees an interesting collaboration with the Free University of Berlin within the CLOUDMAP2 project. CLIWANET and CLOUDMAP2 together will provide an easy accessible complete data set.

The potential use of CLOUDMAP products in the development work in the Nowcasting SAF was discussed. A visiting scientist may work on the verification and (cross) validation of the Nowcasting SAF products within the Initial Operations Phase (IOP). MSG Nowcasting products will be used as a reference in the validation work of the AVHRR SAF products. CLOUDMAP products might be used to "anchor" the reference.

Furthermore, for validation purposes a Swedish ground based network of 3 cameras and an occasional lidar measurement (in Linköping) is available and aircraft reports from the Swedish Airforce can be obtained.

Adam Dybbroe mentioned the fact that the Nowcasting SAF group searches for beta-testers of the software. KNMI showed its interest to become one (possibly in the framework of CLOUDMAP2).

Stefan states that stereo matching using two geostationary satellites is very interesting for nowcasting because of the frequency of the product.

He also thinks that the integrated water droplet product can be valuable in forecasting.

He asks if it is possible to find the top of the stratocumulus with the O2 A band?
The condensation scheme for Cloud Liquid Water (CLW) in data assimilation is not yet workable. The nudging process HIRMES is adjusting the humidity.

Back to main report

Off course KNMI itself is also part of the end-user community, employing operational forecasters, NWP and climate researchers and satellite data experts. Therefore, a last "homeshow" consultation meeting was organised at KNMI. The agenda of the meeting was equivalent to all the other meetings:

It was rather difficult to set a date for a meeting on which all "cloud" experts of KNMI were available. However, a group of 12 experts was brought together and all three end-use target groups were represented in the meeting enabling a lively and interesting discussion.

The main conclusions of the KNMI meeting are:

• Cloud top height (CTH) is certainly not a key parameter in operational forecasting. In NWP this is not yet proven. In climate research CTH is one of the key cloud parameters but a solid error structure of a CTH product is essential for an accepted use.
• The KNMI users (all three end-use groups) have high expectations of the O2 A band enabling the retrieval of microphysical cloud parameters. The inclusion of an O2 A band in future operational satellite missions is very supported.

The following persons attended the roadshow meeting at KNMI. The initials (in alphabetic order) in the first column are used in this discussion report to identify persons who took part in the discussion. The comments made by the individual participants are not verbatim represented in this report and should not be treated as official.