Shared Cost Action Projects in Area 3.3 (CEO) of the Specific Programme for Climate and Environment

HydAlp
Hydrology of Alpine and High Latitude Basins

Project Reference: ENV4-CT96-03634

Internal report RPR2
3-Monthly Progress Report 2: August 1997

Author: H. Rott, S.Quegan, M.Baumgartner, R.Ferguson, D.Miller

Date: 28 August 1997

Institute: Institut für Meteorologie und Geophysik der Universität Innsbruck

This document was produced under the terms and conditions of ContractENV4-CT96-03634for the European Union, DG XXII.

Distribution list

Amendment record

Table of Contents

1. Project summary and management issues *

1.1 Summary of Work Performed *

1.2 Problems and Risks *

1.2.1 Problems *

1.2.2 Risks *

1.3 Project Priorities for Month 7 to 9. *

1.4 Project Deliverables during Project Months 7 to 9 *

1.5 Draft reports *

2. WP100: Project Management *

2.1 WP 110: Technical Co-ordination *

2.1.1 WP 111: Definition of report formats and standards *

2.1.2 WP 112: Definition of software standards *

2.2 WP 120: Management and Administration *

2.2.1 WP 121: Scientific and technical management *

2.2.2 WP 122: Milestones, meetings, risk analysis *

2.3 WP 130: Communication and Interaction with End Customers *

2.3.1 WP 131: Organisation of customer community and meetings *

2.3.2 WP 132: Dissemination of framework standards for customer requirements *

3. WP200: Assessment of customer needs and data base compilation *

3.1.1 WP 201: WP supervision and quality control *

3.2 WP 210: Specification of customer needs *

3.2.1 WP 211: Assessment of customer needs *

3.2.2 WP 212: Customer Workshop *

3.2.3 WP 213: Requirements for Hydropower Management *

3.2.4 WP 214: Interim Report 1 on Customer Requirements *

3.3 WP 220: Hydrological and meteorological data base compilation *

3.3.1 WP 221: HydroMet Data Base for BASAT *

3.3.2 WP 222: HydroMet Data Base for BASCH *

3.3.3 WP 223: HydroMet Data Base for BASSW *

3.3.4 WP 224: HydroMet Data Base for BASUK *

3.3.5 WP 225: Data Access Module *

3.4 WP 230: Remote sensing data base *

3.4.1 WP 231: Search and acquisition of NOAA data *

3.4.2 WP 232: Definition data requirements / acquisition requests *

3.4.3 WP 233: Remote sensing data search *

3.5 WP 240: Field experiments *

3.5.1 WP 241: Preparation and conductance of field work *

4. WP 300: Remote sensing methods and analysis *

4.1 WP 310: Extraction of Basin Characteristics *

4.1.1 WP 311: Review of methods and improvements *

4.1.2 WP 312: Data analysis BASAT *

4.1.3 WP 313: Data analysis BASCH *

4.1.4 WP 314: Data analysis BASUK *

4.1.5 WP 315: Data analysis BASSW *

4.2 WP 320: Remote Sensing Methods for High Resolution Sensors *

4.2.1 WP 321: Improvement of methods for SAR & HROI *

4.2.2 WP 322: Methods for geocoding and information extraction *

4.3 WP 330: Remote Sensing Methods for Medium Resolution Sensors *

4.3.1 WP 331: Review and improvement of methods for MROI *

5. WP 400 Hydrological Modelling *

5.1.1 WP 401: WP Supervision and Quality Control *

5.2 WP 410: Intercomparison SRM/HBV Model *

5.2.1 WP 411: Intercomparison *

6. WP 500 Model Validation and Application *

7. WP 600 Contributions to CEO Enabling Services *

8. Appendix *

8.1 Example of Metadata report (WP310) *

1. Project summary and management issues

• Summary of Work Performed

The main activities in the HydAlp project during project months 4 to 6 included the preparation of data bases, standards, tools for the research activities in hydrology and remote sensing, the assessment of customer needs, and the review and further development of remote sensing and hydrological methods to be used in the project. Three tasks were completed with Final Reports: RI132 – The Framework Standards for Customer Requirements; RI232 – Definition of Data Requirements/Acquisition Requests; RI233 – Remote Sensing Data Search (with one contribution still to be added). The report of Task RI112 (Definition of Software Standards) is under review.

Considerable efforts, involving contributions from all partners, have been devoted to the definition of the concept for data management, including the interfaces with hydrological and remote sensing modules. After written communications, individual partner visits between institutes, and meetings at MLRUI, a draft version of the HYDALP data base and software module system was elaborated during a meeting at IMGI on 13^th August.

A Customer Workshop was held on 24^th and 25^th July 1997 at MLURI to assess the needs for improvements in hydrological modelling and forecasting. A comprehensive report on customer requirements is presently being compiled. This report takes into account the outcome from the workshop, as well as information from other contacts with customers.

Important preparatory activities for the hydrological modelling tasks are the compilation of the hydrometeorological database and the extraction of basin characteristics based on conventional data and remote sensing. These activities are ongoing for all four test basins, though they are in different stages of completeness. Algorithms for snow cover analysis using high resolution sensors have been prepared and are available for testing. As scheduled, the hydrological modelling activities are in an early stage; the intercomparison of the models HBV and RSM, so far based on literature review, has been the only activity. Ahead of the original time schedule, WP600 on HydAlp Contributions to CEO Enabling Services has already been initiated. The presented material includes the HYDALP project overview, interactive administration forms, and information on available data sets.

• Problems and Risks

The main parts of the project activities are progressing well and are on time. Two main problem areas have been identified which so far had only minor impact on the progress of the project, but might in the long run result in major deficiencies.

1. Problems

1. Deficiencies in communication and co-ordination.

Due to the complex interrelationships between the individual tasks of HYDALP a high degree of co-ordination is needed. Though co-ordination improved clearly in comparison to the first three project months, there is still need for improvement. This is obvious from WP reports and from the contributions to the progress report. Different levels of screening and interaction are obvious: Part of the contributions show excellent guidance and co-ordination, whereas others are obviously passed on by the WP manager and WP co-ordinator without any screening.

Action: Guidance and co-ordination for the individual activities should be provided by all responsible scientists at the various levels: WP Manager, WP Co-ordinator, Project Co-ordinator.

The problem of co-ordination is directly related to the communication problem. If technical questions arise or if changes to the planned activities including time scheduling are to be expected, the WP Manager, WP Co-ordinator, and Project Co-ordinator should be informed in time. It should not be necessary to mention that a response is expected if a question on technical or management issues is brought forward. Considering that WP Co-ordinators are supervising around 10 Tasks, and the Project Co-ordinator even more, prompt responses are necessary for proper management of the project.

Action: Inform WP Manager, WP Co-ordinator, and Project Co-ordinator in time on anticipated changes to the activities, including time scheduling. Respond to the technical and management issues brought forward.

• Time delays

Because most of the tasks are dependent on the outcome of other tasks, it is necessary to keep as closely as possible to the time schedule specified in the Technical Annexe to the HYDALP Contract. During months 3 to 6 delays in the review process resulted in time shifts of several weeks for the delivery of the final version of the reports RI 232 and RI233, in spite of the fact that the draft versions had been submitted in time.

Action: The dates for review are known at least several weeks in advance. With proper time management it should be possible to spare the time for review. If a WP Co-ordinator or WP Manager, is not able to take care of the review process, he should inform the Project Co-ordinator and propose another project scientist to take care of the review. (Communication !)

Action: Keep the HydAlp WWW pages up to date with prospective absences and forthcoming meetings

Temporary manpower shortage resulted in delays for compiling one task report (RI213). This problem has been solved. The Project Co-ordinator was informed in time on this problem; therefore it had only minor impacts on other activities.

Short delays for delivery of material for task and progress reports, though not critical for the overall project, make it nevertheless difficult to keep the time schedule at the next higher levels of reporting. Because proper time management requires planning ahead, the person who needs your input should be informed early enough.

Action: Inform the project partner in time on any expected delays. (Communication !)

• Risks

Arising from the problem areas explained above, and from the reports on technical activities, the main risks for the project are related to:

1. Delays in work due to manpower shortage. This is a presently a particular problem of the partner UBE. If the manpower problem cannot be solved soon, major negative impacts are expected for the WP’s related to methods of medium resolution sensors, as well as for intercomparisons of medium/high resolution sensors. It may also cause delays of other WP’s, as already experienced during months 3 to 6.

2. Inability to meet customer needs for hydrological modelling and forecasting in BASUK. As known from the customer survey, the majority of users in this basin wants hydrological forecast for flood prediction. Due to the high temporal dynamics of the snow cover in this region, short time intervals for snow cover monitoring are needed in near real time, which hardly can be obtained by high resolution sensors in an economical way. The use of medium resolution data may be considered as an option to overcome this problem.

• Project Priorities for Month 7 to 9.

• An issue of concern for many activities is the data flow and interface concept. The concept, which has been proposed, should be reviewed by all partners and finalised quickly.

• The compilation of the hydromet data base is due to be finished between month 7 and 9 for the various basins.

• The extraction of physiographic and hydrologic basin characteristics is a complex activity, which should be given increasing attention after the compilation of the hydromet data base has been finished. The output is required for the hydrological modelling tasks.

• The remote sensing methods for high resolution sensors are partially in good shape (regarding snow cover monitoring) but need to be tested in different basins. Part of the work has still to be undertaken (evapotranspiration).

• Few HYDALP-specific activities have been carried out so far on remote sensing methods for medium resolution sensors. These tasks need increasing attention during the next few months.

• In addition to the HBV/SRM intercomparison, which is due to be finished in month 9, major work on hydrological modelling starts in month 7 and is due to be completed in month 9. This entails the extraction and/or calibration of runoff model parameters for the four test basins.

• Project Deliverables during Project Months 7 to 9

The deliverables due in the next three month period are given in the table below:

Where the keys in the deliverable number are given by,

D – Database I – deliverables for internal use of project consortium

R – Report L – deliverables for distribution limited to consortium and customers

RM – Interim Report U - deliverables for unlimited distribution

Database deliverables should be a short report on a completed database, detailing its management system, contents, and capabilities.

The dates of interest for the deliverable review process are given below

These deadlines are supplementary to the 20^th, 24^th and 27^th for monthly task reports, work package reports and work manager reports respectively.

As previously, every three monthly report requires input from WP co-ordinators to the project co-ordinator by the 15^th of months 3, 6, 9.

• Draft reports

All but the final version of reports should include the word DRAFT in the footer and on the first page. The report format (RI111) should also be followed.

Once a final version has been written, a soft copy should be sent by the author to the CEO ftp site. All previous versions should be clearly marked as such, or deleted, to avoid possible confusion.

A hard copy of reports is sent to CEO by IMGI.

• WP100: Project Management

Objective: Project management at scientific, technical, and administrative level; organising the customer community and maintaining the interaction with the customers throughout the project.

Responsible: H. Rott, IMGI

Key issues of scientific and technical management are addressed above.

1. WP 110: Technical Co-ordination

Responsible: R. Caves, SCEOS

Objectives: To define a common format for the reports, as well as quality standards and formats for graphic work (WP 111); to define standards for the software which will be developed within the project; to define data formats and/or protocols for exchange of data files between project partners (WP 112).

WP 111 is complete, and the report RI111 is available

Tasks within WP 110

WP 111 Definition of report formats and standards SCEOS Caves

WP 112 Definition of software standards IMGI Glendinning

• WP 111: Definition of report formats and standards

Responsible: R.Caves, SCEOS, months 1-3

This work package is complete, and available as RI111.

• WP 112: Definition of software standards

Responsible: G.Glendinning, IMGI, month 1-6

The draft report RI112 received OKs from all reviewers, but new ideas and a new general data format were developed at the software meeting at IMGI, on 13^th August.

The General HydAlp Data format is being described and integrated into RI112. This takes the form of header keywords and column data. All files passed to and from user level tools in HydAlp will be passed in this form, to be programmed by DIBAG.

The final report will reference the output report of the Software Meeting for examples of all file types. The final report will be ready at the end of August.

• WP 120: Management and Administration

Responsible: T. Nagler, IMGI

Objectives: Planning and supervision of the project activities in general, administration of the project finances, monitoring of project progress, analysis of potential problems and risks, quality control of deliverables (by the project co-ordinator, in parallel to the WP co-ordinator and WP manager), planning and organisation of meetings.

Tasks within WP 120

WP 121 Scientific and technical management IMGI Nagler

WP 122 Milestones, meetings, risk analysis IMGI Rott

• WP 121: Scientific and technical management

Responsible: T. Nagler, IMGI, month 1-30

Issues and concerns related to this WP are addressed by the project co-ordinator (under Problems-Risks-Responsibilities and Quality Control).

Status of the Action Items from the Kick-Off Meeting:

Status of the Action Items from the 1st Technical Meeting:

Deliverables completed:

RI 232 and RI 233: The delivery of the final version was delayed due to a delay in the review process.

• WP 122: Milestones, meetings, risk analysis

Responsible: H. Rott, IMGI, month 1-30

Scottish Customer Focus group Meeting, MLURI, 22-23.7. 1997. A summary was compiled by J. Morgan Davis and sent to IMGI.

Technical Meeting, MLURI, 24-25.7.1997:: The main topic was the general flowline (including the data access module), which was discussed in detail The flowline integrates remote sensing data, hydrological and meteorological data base management, and hydrological models. Problems regarding the scottish test site BASUK were discussed by the members of MLURI and SCEOS. Participants T. Nagler (IMGI), H. Kleindienst (UBE), R. Caves (SCEOS), O. Turpin (SCEOS), R. Ferguson (SCEOS), G. Wright (MLURI), J. Morgan Davis (MLURI), M. Broatgate (MLURI).

Software Meeting, IMGI, 13 August 97: The general flowline was discussed in detailed. A general HydAlp data format for tabulated data was developed. Each tool of the flowline was discussed in detail. A short report is in compilation describing the general flowline, all the tools which are required, and specific files. Minutes of the meeting were compiled and sent out to the participants. Participants: H. Rott (IMGI), T. Nagler (IMGI), G. Glendinning (IMGI), H. Kleindienst (UBE), Erich Riegler (DIBAG).

For risk analysis, see Section 1.2.2

• WP 130: Communication and Interaction with End Customers

Responsible: G. Wright, MLURI

Objectives: To organise the customers (including project partners and new customers) for the purpose of obtaining specifications of their requirements for hydrological modelling and forecasting, for the assessment of project activities throughout the project, and for assessment of project results.

Tasks within WP 130

WP 131 Organisation of customer community and meetings MLURI Gauld

WP 132 Dissemination of framework standards for customer MLURI Wright
requirements

• WP 131: Organisation of customer community and meetings

Responsible: Gauld, MLURI, month 1-30

This work package is closely linked to WP210, and further information concerning the customer interactions can be found within that work package. Prior to the Scottish Customer Focus Group (SCFG) meeting, some 9 "key" customers were contacted and all responded positively to participation in the SCFG. The main meeting of the SCFG was organised for 22^nd –23^rd July and successfully completed. Since the meeting, a series of follow-up individual meetings and discussions with "key" participants, as well as new organisations identified at the SCFG meeting, have either taken place or are presently being arranged. A more in-depth summary of the initial findings of the SCFG may be found under WP210.

• WP 132: Dissemination of framework standards for customer requirements

Responsible: G. Wright, MLURI, month 1-3

A final copy of this report was produced and distributed at the beginning of Month 4 and this Work Package is now completed.

• WP200: Assessment of customer needs and data base compilation

Responsible: Michael Baumgartner / Hannes Kleindienst, UBE, month 1-24

Objective: Definition of customer needs, preparation (including acquisition) of hydrological, meteorological and remote sensing data, data base compilation, design of an access module

Tasks within WP 200

WP 210 Specification of customer needs MLURI Wright
WP 220 Hydrol. and Meteorol. database compilation UBE Kleindienst
WP 230 Remote sensing database IMGI Glendinning
WP 240 Field experiments IMGI Nagler

• WP 201: WP supervision and quality control

Responsible: H. Kleindienst, UBE, months 1-24

Task 201 has to take care that all deliverables are on time and that problems within WP 200 are going to be solved.

The main activities within this work package are on time, though some activities were delayed due to problems, which have mostly been solved.

The approval procedure of the internal reports RI232 and RI233 was held up at the Work Package co-ordinator level. The final reports were finally issued in mid-April.

The report on task 213 was delayed due to work overload. The draft report has recently been sent out for review.

• WP 210: Specification of customer needs

Responsible: G. Wright, MLURI, months 1-9

The needs for improvements in hydrological modelling and forecasting will be specified in close interaction with customers such as hydropower agencies, hydrological services and environmental agencies.

Tasks within WP 210

WP 211 Assessment of customer needs MLURI Gauld

WP 212 Customer - workshop MLURI Gauld

WP 213 Requirements for hydropower management VERB Pirker

WP 214 Interim report 1 on customer requirements MLURI Wright

For the Customer Focus Group, potential "customers" must have or be prepared to show an active interest in using the main products set out in the HydAlp Project itself. The Scottish Customer Focus Group is formed for the sole purpose of involving "customers" in the specification of their requirements for a hydrological application's run-off model for monitoring and forecasting in alpine and high latitude basins. Primary inputs to this model will be remotely sensed earth observation data and conventional climatic and river basin characteristic information.

• WP 211: Assessment of customer needs

Responsible: Gauld, MLURI, months 3-7

This part of the HydAlp project is closely linked with the outcome of WP130 and WP212. Almost all of the Scottish Customer Focus Group participants not attending the Workshop have been contacted and discussion meetings were arranged for the 20th & 26th August. Further suggested participants have been carefully screened for direct interest in the role and objectives of the HydAlp project and will be followed up by use of the SCFG Questionnaire or by interactive discussion. It was found that initial suggestions of customer requirements, from SEPA (Scottish Environmental Protection Agency), were closely reflected by the results of the SCFG Workshop meeting (detailed in WP212).

• WP 212: Customer Workshop

Responsible: Gauld: MLURI, months 3-6

The Scottish Customer Focus Group Workshop took place at MLURI on 22^nd and 23^rd July 1997. The response from initial SCFG participants to the SCFG Workshop has been very positive with a good complement of organisations represented. It was unfortunate that a conflicting timetable meant that a "Key" participant, Scottish Hydro-Electric, was not able to attend, but their contribution will still be included in the final report through the application of the second phase of the SCFG - the informal interactive discussion with individual customers.

The draft report for RI212 has been received.

Contributing Participants (by Organisation):

Dr. Ken Pugh & Mr. Nigel Goody (SEPA), Mr. Les Street (RSPB), Dr. Johnothan Stacy (SNH), Mr. David Crichton (General Accident Insurance), Mr. Dan Woodrow (Anite Systems), Mrs. Marian Austin (Scottish Ski Federation), Mr. Otto Pirker (Austrian Hydropower), Ms. Rachel Helliwell (MLURI Hydrologist).

Potential customers were invited to comment on their particular requirements for a snowmelt and runoff model and it was determined, at the outset, that flood risk assessment represented the principle concern of all participants.

The customers' "wish-list" of needs is summarised below. It is important to note that customers were also keen to have relative costs for each data source detailed and highlighted the lack of concerted financial resources currently available to address the entire topic of flooding.

NB: Output potential is not specifically the modelling potential within HydAlp but within future operational production.

Action: It was agreed at the SCFG Workshop meeting by the modellers present that they should report back to the SCFG whether the needs specified by the customers could be addressed within the HydAlp project. The modellers' response to each "requirement" should be in the format:

i) Can be addressed within HydAlp,

ii) Could possibly be addressed,

iii) impossible to address within the scope of HydAlp.

Conclusions reached by the SCFG Workshop were as follows:

• All customers stressed the complex inter-relationships of environmental factors when used to predict floods.

• They were also aware of the considerable benefits in modelling flood predictions providing local environmental factors could be incorporated within the sophisticated models.

• Further refinement of specifications is needed by involving additional customers (particularly from Switzerland, Sweden and Austria).

• WP 213: Requirements for Hydropower Management

Responsible: Pirker: VERB, months 1-4

Due to temporary shortage of manpower the work was delayed. The draft report, RI213, is now in review.

• WP 214: Interim Report 1 on Customer Requirements

Responsible: G. Wright, MLURI, months 7-9.

This part of Work Package 210 does not need to be actioned yet.

• WP 220: Hydrological and meteorological data base compilation

Responsible: H. Kleindienst, UBE, months 2-9

Tasks within WP 220

WP 221 Hydromet database for BASAT IMGI Glendinning

WP 222 Hydromet database for BASCH UBE Kleindienst

WP 223 Hydromet database for BASSW SMHI Johansson

WP 224 Hydromet database for BASUK MLURI Morgan-Davies

WP 225 Access module design UBE Kleindienst

Within this Work Package the hydrological and meteorological data will be compiled and set up to be used for hydrological modelling.

This report gives an overview of the data, which is already available at the institutes. It also defines the main periods of interest, which will be hydrologically simulated.

• WP 221: HydroMet Data Base for BASAT

Responsible: G. Glendinning: IMGI, months 2-7

Work has continued collating data for BASAT, and it has been attempted to extend the database to include more high-resolution data. A further meteorological station is proposed to complement those already present.

Liaisons with Tauern Kraftwerk (TKW) have almost completed the data search at present.

Hydrological data is available from Persal, Wasserfall, and Tuxbach Überleitung.

Meteorological data is available from Mayrhofen, Schlegeis, Plattkopf and Stillup.

Hydrological and meteorological data

The following table sets out the hydromet data available for BASAT. The data is stored in simple ASCII files: each file represented by a row in the table.

Table 1: Available HydroMet data for BASAT – T represents air temperature (07, 14 & 19 being hours in the day), ppt is precipitation, Q is runoff and Z is snow depth (Znew is newly fallen snow depth), avg is daily average.

Time period to be simulated

The modelling will use previous year's data to set up model parameters, and run on data from the melt seasons 1996, 1997, 1998 and 1999 (with real time data). The melt season runs from April to September for BASAT.

Data outstanding

1981 -1996 daily meteorological data may be available for Gschösswand, Lanersbach, Ginzling, Innerschmirn – Obern and Wattener Lizum. These will be obtained through the Hydrographische Dienst, Österreich. Unfortunately, they are only available as hand-written records, so will not be analysed in their entirety.

There are plans to place a further meteorological station in Tuxbach in autumn 1997 at around 2000m a.s.l.

1997 data will be processed when available.

Conclusion

Sufficient hydromet data is now available to calibrate and validate runoff models for BASAT up to 1996. The further data will improve the accuracy of the meteorological model input. The data is all yet to be stored in the required format for HydAlp.

• WP 222: HydroMet Data Base for BASCH

Responsible: H. Kleindienst, UBE, months 2-7

During the last three months a complete set of meteorological data from seven stations was compiled and organised in a database. Further on hydrological data for several ablation periods were as well stored in a database.

Some tests were carried to check if the cloud cover data from the meteo stations could be used to derive dates for which cloud-free images would be available. Although no concrete procedure can be proposed yet the cloud cover seems to be useful information.

Hydrological and meteorological data

The following table shows the data sets available at UBE.

Table 2: Available HydroMet data for BASCH – T represents air temperature (7, 14 & 19 being hours in the day), P is precipitation, B is Cloud cover, rH is relative humidity, Q is runoff (corr is corrected, nat is natural) and Z is snow depth (Znew is newly fallen snow depth), avg is daily average.

The meteorological data is provided and verified by SMA (Swiss meteo office). It therefore can be assumed that the data is of good quality. The quality of the hydrological data cannot be assessed.

Time period to be simulated

Most of the periods for which hydrological data is available are already simulated with the SRM. The simulations will be repeated (in order to eventually improve model results) and if possible new data sets will be applied.

Data outstanding

It is the aim to get additional hydrological data sets, if possible for more recent years (1994-96). Therefore several hydropower companies must be contacted in order to get the data of water abstractions.

Several sources of forecasted meteorological data have been checked (SMA, private meteo office, newspaper). It is planned to set up a database of forecasted meteorological data as a basis to test operational runoff and snow cover forecast.

It is planned to get additional data of snow cover conditions from the Snow and Avalanche Research Institute in Davos (SLF-WSL). However, there has been no contact yet.

Conclusion

There is sufficient data available to run both models. Some more recent data will be obtained. A test of operational runoff forecast is planned.

• WP 223: HydroMet Data Base for BASSW

Responsible: B. Johannson, SMHI, months 2-7

We have here at the institute, in our databases, the necessary hydromet data to run the model. A few of the meteorological stations are synoptic i.e. we receive data on a real time basis. In a forecasting situation, it will also be possible to get runoff data in real time. We may also arrange with some observers at the non-synoptic meteorological stations to phone them to get precipitation data for the last days before a forecast.

We have conversion programs to extract data from our databases, and create the PTQW file used by the HBV model, but of course we need to do some programming to convert data to the transfer format the HYDALP database will require.

We will use the following meteorological stations:

16798 Kvikkjokk 66 degrees 57' 17 degrees 44' 337m, synoptic station, 1959-

16897 Tjamotis 66 degrees 55' 18 degrees 32' 300m. precipitation, 1920-

16988 Jokkmokk 66 degrees 37' 19 degrees 38' 260m, synoptic station, 1961-

17792 Ritsem 67 degrees 43' 17 degrees 28' 521m, synoptic station, 1981-

17879 Aluokta 67 degrees 18' 18 degrees 54' 385m, precipitation + temperature,

Runoff data and water stage are taken from:

1969 Tjaktjajaure, outflow 1974-1994

1966 Tjaktjajaure, water stage 1974-1994

40071 Tjaktjajaure, inflow 1995-

Long-term monthly mean values of potential evapotranspiration have been computed for the station Kvikkjokk (16798).

Unfortunately none of the stations is within the catchment, so the remote sensing data may really improve the simulations!

• WP 224: HydroMet Data Base for BASUK

Responsible: J. Morgan-Davies, MLURI, months 2-9

All available Hydromet data has now been sourced and is presently being acquired or compiled. Summer vacations or busy schedules of organisations and the presence of SCFG meetings has meant that data retrieval from outside sources (non MLURI) have not been forthcoming. Meetings for finalisation of data and acquisition have been planned for the next month (7 of project).

Most data is expected to be made available from the UK Meteorological Office archives. SEPA (who will also provide riverflow data to complement other data types if required) will also be providing "fill-in" data. All data will be compiled at MLURI in Ascii format from ORACLE where the MORECS met-data archive is held.

All available meteorological stations will be included in the database. Aviemore is the nearest and most logical weather station to contribute data to the Feshie study site. It does not lie directly in the Feshie sub-catchment but in the overall Upper Spey catchment. Daily rainfall, mean temperature, potential and actual evapotranspiration figures should all be available through Met office sources.

• WP 225: Data Access Module

Responsible: H. Kleindienst, UBE, months 2-9

During the month 4 and 5 of the project the use and a possible design of a data access module were widely discussed via e-mail. At the technical meeting in Aberdeen (24/25 July 97) a version of the access module, as part of a general Hydalp-Flowline, was discussed.

At an additional software meeting in Innsbruck (13 Aug 97) the design of a general project flowline was specified and a final version of the data access module was set up.

Both will be distributed to all project partners for final comments at the end of August 1997, and documented in interim reports.

• WP 230: Remote sensing data base

Responsible: G. Glendinning, IMGI, months 1-24

Work Package 230, Remote Sensing Data Base, comprises the project requirements for selecting and ordering remote sensing data in HYDALP. For this process, the details of basin locations, data already available and further requirements are needed.

Tasks within WP 230

WP 231 Search and acquisition of NOAA data UBE Ottersberg
WP 232 Definition data requirements / acquisition requests IMGI Nagler
WP 233 Remote sensing data search SCEOS Turpin/Nagler

• WP 231: Search and acquisition of NOAA data

Responsible: R. Ottersberg: UBE , months 1-24

This work is ongoing. Receiving stations and archives are OK.

Month 4: 95% of data was received and stored between 20.05.97 and 19.6.97
Month 5: 96% of data was received and stored between 20.06.97 and 19.7.97

A complete list of the recorded images is available at:

!!NEW!! http://saturn.unibe.ch (select ‘Satellite receiving station’) !!NEW!!

• WP 232: Definition data requirements / acquisition requests

Responsible: T. Nagler: IMGI, months 1-4

The report RI232 was sent for review on 23^rd June 1997, and the three reviews incorporated by the 30^th June 1997. There was then a 6 week delay before the final OK by WP co-ordinator was given. The report has been distributed.

• WP 233: Remote sensing data search

Responsible: O. Turpin: SCEOS, months 1-4

The report RI233 was sent for review on 18^th June 1997. It required contributions from IMGI and UBE. The IMGI contribution was incorporated by 3^rd July 1997, and the corrections passed by WP manager and project co-ordinator on 7^th July 1997. There was then a 5 week delay before the OK by WP co-ordinator was given. The report has been distributed , with the UBE contribution is still outstanding.

• WP 240: Field experiments

Responsible: T. Nagler, IMGI, months 1-4, 13-16

Comparative field experiments will be carried out in the Austrian and Scottish test basins for the intercomparison of remote sensing techniques and the assessment of the accuracy and applicability of the data analysis methods.

Only Task 241 was active, during month 4 (June 1997).

Tasks within WP 240

Task 241 Preparation and conductance of Field Work IMGI Nagler
Task 242 Preparation and conductance of Field Work MLURI Bell
Task 243 Field work SCEOS Turpin
Task 244 Field work UBE Kleindienst

• WP 241: Preparation and conductance of field work

Responsible: T. Nagler, IMGI, months 1-4, 14-16

The test site Tuxbach was explored to find appropriate sites for collecting snow data (Hobalm, Grieralm and /or Höllenstein Hütte, Kleegrube). Further information on snow pit data was retrieved from the avalanche warning office of Tyrol (Lawinenwarndienst) of the Tauern-Kraft-Werke (TKW) / Verbund in Mayrhofen (Mr. H. Rass). The avalanche office collects snow pit data at 5 locations, two of which are in the test site (Penken, 2095 m, Wanglspitz, 2420 m), in a cycle of 2 weeks during the winter season. For the winter 1996/97 the parameters stratigraphy, snow depth, density, grain size, and hardness are summarised in an internal report of the TKW. Snow pit data of previous years are available at the avalanche office of the TKW.

Visits to the runoff gauges Tuxbach Überleitung, Wasserfall, and Persal and photographic documentation of the ablation patterns of glaciated areas are planned for September 1997. The purpose of these visits is to prepare for the planned hydrological work in spring 1998.

• WP 300: Remote sensing methods and analysis

Responsible: Shaun Quegan, SCEOS, months 1-24

Objective: This work package has the following objectives:

· To review available methods for remote sensing data analysis, to identify needs for improvements regarding the application in hydrological models, and to implement the methodological improvements.

· To extract hydrological relevant information from the remote sensing data to be used as input for hydrological modelling and forecasting.

Optical sensors are to be used for mapping areal extent of snow cover, surface albedo, and land surface types, while SAR is to be used for mapping the extent of melting snow cover. Possibilities for estimating evapotranspiration will also be investigated. The work is broken down into four sub-work packages:

WP 310 Extraction of Basin Characteristics. MLURI, Miller

WP 320 Remote Sensing Methods for High Resolution Sensors. IMGI, Nagler

WP 330 Remote Sensing Methods for Medium Resolution Sensors. UBE, Baumgartner

WP 340 Earth Observation Data Analysis. SCEOS, Caves

Work on WPs 310 and 320 is progressing well. Due to the absence of M Baumgartner progress on WP 330 has been slow. Work on WP 340 does not start until month twelve. The previous 3-monthly progress report highlighted the need for increased co-ordination and communication between individual tasks and work packages.

R Caves’ visit to IMGI (27th June - 4th July), and the technical meeting at MLURI (24-25th July) following the SCFG have helped focus WP 320 work, in particular the definition of flow lines and interface to other WPs. However, the choice of SAR geocoding solution and HROI radiometric correction routines have yet to be finalised, and the overall flow line implemented.

Communication within and external to WP 310 had been slow though is now much improved, partially due to the compilation of metadata at MLURI and the preparation of this report. The need for RL311 "Review of methods and improvements for extracting basin characteristics" to address generic issues has been stressed.

A good overview of the status of the MROI work was gained from R Caves’ visit to UBE (25-26th June). However, no further progress has been reported from UBE since. As a result of this the integration of WP 330 with other WPs is much further behind that of WPs 310 and 320. The need for this integration is highlighted by the fact that WP 300 is meant to include an intercomparison of medium and high resolution methods. It should also be noted that in terms of information content, availability of data and accuracy, any intercomparison should be conducted in the same basin. In principle, any of the three larger basins (BASCH, BASSW and BASUK) would be suitable for this intercomparison, as MROI methods are only applicable to larger basins. BASCH (3249 km²) or sub-basins would be a good choice, because methodological developments for medium resolution sensors will be done with this basin. For verification of these methods, the intercomp arison of high resolution data would be of good value. IMGI will assist UBE with the analysis of the high resolution data. BASSW is the next largest basin (2250 km²) and the expertise exists at SCEOS for analysing both medium and high resolution datasets. For BASUK (1268 km²), there is more uncertainty regarding the applicability of the hydrological models. The upshot of the SCFG meeting, however, shows a need for close time sequences of data in BASUK which points towards MROI. BASAT is too small (130 km²).

The choice of a basin for this intercomparison will be addressed in the progress meeting (17-18^th November 1997).

• WP 310: Extraction of Basin Characteristics

Responsible: D. Miller, MLURI, months 1-12

Objectives: To extract the physiographic characteristics of each test basin. These data will be compiled into a Geographic Information System (GIS) for integration with other datasets (hydrological, climatological, Earth Observation (EO), and ground based) in hydrological modelling. The minimum characteristics that are required are those that will provide the models with their key inputs, such as altitudinal bands and selected land cover classes. These data will be gathered for entire basins and in greater detail for selected sub-basins to allow for a more comprehensive demonstration of the potential of the models.

Tasks within WP 310

WP 311 Review of methods and improvements MLURI Miller

WP 312 Data analysis for basin BASAT IMGI Glendinning

WP 313 Data analysis for basin BASCH UBE Baumgartner

WP 314 Data analysis for basin BASUK MLURI Miller

WP 315 Data analysis for basin BASSW SCEOS Turpin

General co-ordination (internal and external).

Tasks 312, 313, 314 and 315 became active during months four and five. The objective of these Tasks is to extract the basin characteristics required for the running of the hydrological models in each of the test basins. These characteristics include land cover, topography represented as a Digital Elevation Model (DEM), derivatives of topography such as slope or aspect, and mapped surface hydrological features (such as basin and sub-basin boundaries and directions of flow).

No significant problems have been encountered with the data collection or preliminary analysis for any of the basins, although progress is still at an early stage.

Work within WP 310 has been closely linked with that in:

• WP 132 in liaison with the Scottish Customer Focus Group, specifically the Scottish Environmental Protection Agency who chair the group and are themselves, potentially, a principal user with whom the BASUK boundary was (provisionally) agreed;

• WPs 220 and 320 on how information relating to basin characteristics will be integrated with information extracted from RS data for input to the hydrological models. This was a major item of discussion at the MLURI technical meeting following the SCFG;

• WP 600 for the development of the meta data pages;

• Initial discussion is taking place with WP 302 and WP 321 on the division of content of the Reports on the `Review of Methods' in each WP where there is potential for overlap or duplication.

Internal links are being actively pursued now that data are becoming available for each basin and issues of documentation and procedure are current. Prior to this stage such links were not established in great depth.

Meta data:

Each dataset will be comprehensively described within a standard metadata form. An example is available in the Appendix. This provides details on the origins, content, reliability, geometric properties and copyright/access to the datasets. This aspect of the WP has been developed together with WP 600.

The technical content of the meta data includes:

• data format (raster or vector)

• software used in the derivation of the data

• file format of datasets and examples

• projection system and associated information

• minimum mapping unit and measurable distance

• attribute codes

• methods of data derivation

• reliability and accuracy statements

• WP 311: Review of methods and improvements

Responsible: D. Miller, MLURI, months 1-8

The structure and content of the report RL311 has been formulated with the WP Co-ordinator. The report will review generic methods for extracting basin characteristics, and will be targeted at potential users of the results of HydAlp who wish to apply the procedures developed to a basin not included within the project. The proposed structure for the report is contained in the Appendix.

Issues relating to the use of RS data for extracting/updating land cover information will be dealt with in so far as they are specific to this WP, e.g. choice of data and land cover classification routines. It is likely that geocoding issues will be sufficiently covered by WP 320 and 330. Radiometric correction will also be largely covered by these WPs.

Section headings for RL311

Introduction

- what we need, why we need it, how accurate it must be

- definitions and terminology

-- basins/catchments, characteristics

-- land cover, elevation (and associated), hydrology

Characteristics (descriptions, reliability, availability)

- land cover

- elevation

- hydrology

Methods

-- land cover

-- elevation

-- hydrology

-- combined units (e.g. units based upon combinations of physiographic data such as soil, land cover and landform)

will be dealt with under the following headings

--- data sources (origin, history)

--- data types

--- data structures (raster, vector, object linked)

--- derivation

---- measurement (field, indirect observation from imagery, photography or maps)

---- modelled (discussion of different types of modelling and analytical solutions with respect to catchment characteristics; e.g. alternative models of elevation, inferred land cover types, network models of hydrology)

---- accuracy (discussion of absolute, relative, scale and attribute errors)

---- coding (interpretation coding and links to accuracy, automated classifications, nature of coding systems for polygon, linear and point data and raster data)

---- currency (issues to be considered with respect to length of time for which data is valid)

Metadata

- definition

- "value"

- use

- example

Conclusions

References

• WP 312: Data analysis BASAT

Responsible: Graham Glendinning, IMGI, months 4-11

1. DEM

The DEM covering BASAT extends over all of Zillertal, Tirol.

Source, Format and Copyright:

The DEM was derived from geodetic maps. It is supplied as raster data from a vector source. It is stored and manipulated in PCI EASI/PACE 6.1 in *.PIX format. A supplemental digitised map (an amalgamation of OEK 50 digitised maps scanned at 400dpi) is utilised to aid GCP selection for geocoding. Use of the DEM is restricted to IMGI personnel.

Characteristics:

Slope and aspect information have been derived from the DEM.

2. Land Cover

Land cover is derived solely from satellite imagery. Landsat Thematic Mapper data were utilised, from 16^th August 1992 and 17^th September 1992. These images were geocoded in EASI / PACE (same projection specifications as for the DEM). Planetary albedo maps were derived, and simple thresholding techniques applied for a preliminary class derivation. Classes derived include:

• Glacier (TM 3/TM 5 > 3.5)

• Forest (TM 3/TM 4 <=0.28)

• Grass and low vegetation(TM 3/TM 4 >= 0.28 and < 0.65)

• Rocks and moraine (TM 3/TM 4 >=0.65)

Surface albedo will be required for more accurate or complex class calculations, and aspect and slope will also be introduced to the calculation. Aspect and slope have been derived from the DEM within PCI.

The basin extent was calculated using the watershed routines of PCI EASI/PACE. The downstream extent was limited by the lower stream gauge, Persal.

Meta data has been produced for BASAT extent and the DEM, but not yet for Land Cover.

• WP 313: Data analysis BASCH

Responsible: Michael Baumgartner, UBE, months 4-11

Due to the absence of the Task manager no update was available for BASCH, neither was a monthly report supplied for July.

The below information is compiled by Ron Caves on behalf of the WP manager.

The June monthly report stated that data is available for:

• Basin extent.

• Hydrological features - natural and artificial lakes, and main rivers.

• Meteorological stations.

• Elevation zones with an equidistance of 500m.

The above data is to be transferred into the ArcView GIS. A 1km DEM with 100m height accuracy is also available. This is sufficient for analysing MROI data.

• WP 314: Data analysis BASUK

Responsible: David Miller, MLURI, months 5-12

Some discussion has been undertaken on the selection of the final definition of each basin and in the case of BASUK a revision of the extent of the basin is currently being discussed. The mapped, hydrological data for the area has been ordered from the Ordnance Survey together with the 1:10 000 vector contours for the entire basin.

Geographic data that is available to HydAlp for the BASUK is listed in Appendix 3.

Digital Elevation Models

The 1:50000 raster DEM from Ordnance Survey will be used as the principal source of elevation data. The DEM has been resampled to 25m and 12.5m using bilinear interpolation for compatibility with the output resolution of the rectified satellite imagery. A raster DEM will be derived from the 1:10000 contour data.

Two tests are proposed using different source scales of elevation data:

• On the quality of the rectification of the satellite imagery.

• The sensitivity of the hydrological models to alternative inputs scales. These tests are to establish guidelines on the `fit for purpose' of the input data, the significance of which is to present to users the implications of the type, source and quality of the input data.

Two preliminary studies are been carried out:

• On the sensitivity of the automated delimitation of the basins to the scale of the input data, with respect to the location of the boundary and the summary statistics for characteristics for each sub-basin.

• On the sensitivity of the derived values of slope and aspect to the algorithm used.

The 5m DEM for the Glen Feshie sub-basin will be completed by the end of August 1997 and this will provide a large scale input dataset for the study of the sub-basin.

Land Cover

The land cover data has been recoded to a preliminary classification of

• Woodland

• Grassland

• Wetland

• Bare rock

• Lochs (lakes)

• Urban areas

Significant further discussion will be required on the detailed reclassification and it is anticipated that there will be several versions of the classification to permit alternative scenarios of model runs to be undertaken.

Hydrology

Two sources are being used. 1. Ordnance Survey 1:10000 Landline data for Glen Feshie and the "larger" rivers within BASUK; 2. "Main" hydrological features derived from the elevation models. The outputs from source 2 will be calibrated by reference to source 1.

Geographic data available to BASUK (WP314)

Digital Elevation Models (DEMs)

Land Cover

Hydrology (rivers, streams, and lochs)

Loch boundaries

Catchment boundaries

• WP 315: Data analysis BASSW

Responsible: Owen Turpin, SCEOS, months 4-11

The following data is available for BASSW.

DEM in ARC/INFO ASCIIGRID format.

Catchment boundaries for Tjaktjajaure and Litnok, in ARC/INFO export ASCII-code and in ARC/INFO ungenerate ASCII-code formats.

The map projection is RT90 with co-ordinate system:

and parameters:

The DEM has been supplied to SCEOS but analysis has not yet commenced. The DEM may only be used by SCEOS for the purposes agreed with SMHI.

The following basin characteristics will be used for modelling:

• Land cover: forested/unforested, lake and glacier;

• Hypsograph for each sub-basin.

All data except glacier area are already mapped in digital form on a 500 x 500m grid.

• WP 320: Remote Sensing Methods for High Resolution Sensors

Responsible: Thomas Nagler, IMGI, months 2-12

Objectives: To assess the available methods for the extraction of information from high resolution optical and radar imaging sensors for mapping extent and properties of the snow pack and for estimation of evapotranspiration. To identify deficiencies, and to implement modifications regarding the use in hydrological models.

Tasks within WP 320

WP 321 Improvement of methods for SAR & HROI IMGI Nagler

WP 322 Methods for geocoding and information extraction SCEOS Caves

Status of WP 320

This WP requires close co-ordination between IMGI and SCEOS. The required work has been well defined within the last few months, including the definition of general flow-lines, search for available software, definition of required programs, and formats for providing remote sensing products to other tasks. WP 320 is within the time schedule.

Visit by R Caves to IMGI: 27 June to 4 July 1997

This visit was the main focus of WP 320 over the last three months. The main purpose of the visit was to co-ordinate the work and to discuss software development within WP 320 (R Caves has compiled an internal SCEOS report on this visit and his visit to GIUB). Based on procedures developed by Nagler (1996) the flow lines for processing SAR and HROI data were defined in detail. To take advantage of standard display, visualisation, and image processing tools it was decided to develop the analysis software within the PCI EASI/PACE UNIX environment, which all relevant HydAlp partners have access to. It was agreed that the processing approach should be modular so that modules can be upgraded, changed, added or deleted as desired.

All relevant external EASI/PACE routines developed by IMGI for SAR snow mapping and HROI surface classification were demonstrated and copies were supplied for installing at SCEOS; these programs need to be modified to meet HydAlp software and documentation standards (RI112). Compilation problems were initially encountered when these routines were installed at SCEOS. Also, a license for the PCI Software Toolbox had to be purchased before the routines could be run; a check was made to ensure that SCEOS had a license for all other PCI modules used by IMGI. A number of new programs, which are required for this WP, were also defined. Strict version control procedures will be required for routines jointly developed between IMGI and SCEOS. A Remote Sensing Product File (RSPF) format was proposed for integrating the information extracted from RS data (e.g. snow coverage) into the hydrological models.

For testing the procedures, a set of images is required;

BASAT: images available at IMGI

BASUK and BASSW: no images are available at SCEOS.

1. WP 321: Improvement of methods for SAR & HROI

Responsible: T. Nagler, IMGI, months 2-12

The processing of SAR and HROI will be developed in modules, called RS.SAR and RS.HROI. These modules are further separated into sub-modules, labelled with the extension 1, 2, etc.

SAR Image Analysis Module - RS.SAR

The generation of snow maps from SAR data relies on the procedure proposed by Nagler (1996). Figure 3.1 shows the flow chart of the main processing steps. It can be divided into two modules RS.SAR.1 and RS.SAR.2, both of which comprise several sub-modules. The SAR flow line is based on ERS PRI images. Modifications will be required for Radarsat.

The module RS.SAR.1 deals with the generation of a stack of co-registered repeat-pass SAR images. For each pass a geometrical master image (GMI) is selected. Due to the use of repeat pass images full geocoding need only be defined for the GMI of each pass. Geocoding routines are described under WP 322. Each image is absolutely calibrated by applying SAR-processor dependent calibration factors. A routine for calibration of ERS-1/2 PRI data is available at IMGI including replica power correction, range spreading loss, and application of the improved antenna pattern (program PRIADJ). A corresponding calibration routine is required for Radarsat data. Repeat pass images are co-registered to the corresponding GMI using coarse and fine matching. The Coarse Matcher determines the linear shift in azimuth and range between the actual SAR image and the GMI (matching will be done in pixel scale). If significant variations in the heading and look angle are observed the actual im age is resampled to the GMI based on the cross-correlation of image chips (Fine Matcher, which includes rotation and stretching capabilities). Both matching routines will be developed in co-operation with Ron Caves, SCEOS. The co-registered images including meta-information will be catalogued.

The module RS.SAR.2 deals with the generation of the extent of wet snow for a single period. Crossing overflights (ascending and descending passes within several days) will be used to reduce the loss of information due to layover/shadow. The procedure is described in detail in Nagler (1996). For generation of a single snow map four SAR scenes are required in total: 2 SAR snow images (SI), which are acquired when the snow extent is to be determined, 2 SAR reference images (RI), acquired during absolutely dry snow or snow free conditions. The same SAR reference images will be used for all dates. Prior to rationing, speckle smoothing must be applied (e.g. using the EASI/PACE filter ffrost). The IMGI routine snwmap is then used for forming the ratio SI/RI and for generating the snow extent bitmap. The presence of wetland areas was identified as a problem when selecting reference images for BASUK and possibly BASSW.

The geocoded ratio images of ascending and descending passes are combined to reduce the loss of information due to layover (snwmer). So long as geocoding is accurate the ascending/descending combination is a relatively straightforward task based on the incidence angle maps calculated from orbit information and the DEM. It should be noted that the time lag between ascending and descending passes varies markedly between test basins. For BASAT the lag is only half a day but for BASUK and BASSW it is one and half days and six and a half days respectively. These longer lags will permit more change to occur between passes and hence reduce the validity of the ascending/descending combination. The snow cover map, the residual layover/shadow map, and meta-information describing all processing steps (software version, etc.) will be listed in a catalogue.

HROI Data Analysis - RS.HROI

HROI data will be used for classification of surface types, snow mapping, and estimation of evaporation. Figure 3.2 shows the flow-line for surface classification and snow mapping, which is based on a procedure by Nagler (1996). The HROI flow-line is based on the use of Landsat TM data and will need to be modified for SPOT and other HROI sensors. The computer code for derivation of the planetary albedo (albedo determined at the sensor, without any atmospheric correction, program tmpa) and the surface albedo (requires atmospheric transfer model, program tmsa) from Landsat-5 Thematic Mapper data were developed at IMGI and run under the EASI/PACE environment. The geocoding requires the routines smodel and sortho, which are part of the EASI/PACE Ortho-Rectification Module. The classification is based on a hierarchical decision tree applied on maps of the planetary albedo and surface albedo.

Several modifications and improvements of the existing software are required. This includes the calibration routine, where the spectral channel characteristics of other sensors (e.g. SPOT, LISS-III, etc.) have to be specified. Problems in the surface albedo calculation occur mainly in shadowed areas and in steep valleys (with values which are significantly too high). To correct for the impact of the reflected solar irradiation from surrounding slopes the recent publications on this topic will be studied (Dozier and Frew, 1990, Sandmeier and Itten, 1997). Improvements in the atmospheric radiative transfer calculations will be investigated by using the 6S (Second Simulation of the Satellite Signal in the Solar Spectrum, University of Lille) (Vermonte et al., 1997).

Remote Sensing Product File (RSPF)

The RSPF contains the information extracted from the remote sensing data (e.g. snow cover fraction) for sub-areas specified by various topographic parameters (sub-basin, elevation, aspect, slope) and thematic parameters (surface classes). The high resolution (HR) RSPF will use sub-areas with the finest sub-division of these parameters (elevation steps of 10 m, 16 aspect classes, 8 slope classes, various surface classes) and will be combined to get the remote sensing product per Hydrological Unit. Methods for integrating RSPF files with HydroMet data in the hydrological models are being discussed with the modellers. This was a major item of discussion at the recent MLURI technical meeting following the SCFG.

References

Dozier J. and J. Frew, Rapid Calculation of terrain parameters for radiation modelling from digital elevation data, IEEE Trans. Geosci. Remote Sensing, Vol. 28, No. 5, pp. 963-969, 1990.

Nagler T. , Methods and Analysis of Synthetic Aperture Radar Data from ERS-1 and X-SAR for Snow and Glacier Applications, PhD Thesis, University of Innsbruck, 1996.

Sandmeier S. and K. I. Itten, A physically-based model to correct atmospheric and illumination effects in optical satellite data of rugged terrain, IEEE Geosci Remote Sensing, Vol. 35, No.3 May, 1997..

Vermonte E.F., D. Tanré, J.L. Deuzé, M. Herman, and J.J. Morcrette, Second Simulation of the Satellite Signal in the Solar Spectrum, 6S: An Overview. IEEE Geosci Remote Sensing, Vol. 35, No.3 May, 1997.

• WP 322: Methods for geocoding and information extraction

Responsible: R. Caves, SCEOS, months 2-10

The basis for our recommended form of geocoding within HYDALP needs to be clarified. R. Caves will compile a discussion document consisting of a list of the issues that need to be addressed. Comments on these issues would be requested from IMGI, GIUB and SCEOS, and their replies would form the basis for defining clear criteria for choosing a geocoding solution.

This document needs swift attention if it is to have any impact on our recommendations. T.Nagler has suggested that a short comparison of the advantages/disadvantages and of required parameters of the two available geocoding software packages (SCEOS and IMGI: EASI/PACE, only IMGI: RSG) would be useful; one main problem will be the detection of accurate ground control points.

SAR Geocoding

Accurate geocoding of SAR images is required for combination of snow maps of ascending and descending orbits. The accuracy of the geocoding depends on the quality of: the digital elevation model, the sensor and orbit parameters, the ground control points, and the geocoding algorithm used.

Two geocoding solutions are being discussed:

• Matching the real image to a simulated image generated from the DEM and then resampling the real image back to the DEM projection by inverting the process used to generate the simulated image. This approach utilises the PCI sarsim1 (SAR image simulation), sargeo (ortho-rectification), and sarsim2 (auxiliary products like layover, shadow mask) routines.

• The exact range doppler solution using the RSG (Remote Sensing Graz) routines.

IMGI have much more experience of using the image simulation and matching approach than the range doppler approach. With the latter, problems were encountered in identifying suitable GCPs in mountainous areas. Only the PCI routines are available at both IMGI and SCEOS. Any decision on whether to test the RSG software should wait until the discussion document on SAR geocoding had been compiled, circulated and responses analysed. RSG is marketed in the UK by ERDAS. Erich Riegler may be able to help obtain a copy of the RSG at a reduced rate.

In EASI/PACE image simulation (sarsim1) is based on the following parameters:

• satellite altitude

• satellite heading

• look angle to the first pixel

• pixel spacing

All parameters should be available from the image header. Fine tuning of the heading, incidence angle and pixel spacing is used to improve the simulation relative to the real image. Little effect is noted from changing the altitude. S Quegan noted that fine tuning of sarsim1 parameters is mostly based on heuristics. He considered that objectivity and repeatability should be one of the topics considered in the discussion document on geocoding. Matching of the two images is based on layover features. The IMGI/PCI tempmat routine is used to match layover features in the real and simulated images.

HROI Geocoding

IMGI have suggested that the PCI routines smodel, sortho (PCI ortho-rectification package) be used initially. These routines have been tested successfully by IMGI for BASAT. The EASI/PACE-geocoding routines require a digital elevation model, sensor and orbit parameter (extracted from original data files) and ground control points. The sensor and orbit parameter are extracted automatically by using EASI/PACE reading routines, available for Landsat-5 TM (EASI/PACE-program cdlandc) and SPOT images (EASI/PACE-program cdspot). For collecting ground control points the EASI/PACE-program gcpworks is used; as the geo-referenced image a digital copy of the topographic map or a digitizer is required.

Geocoding of HROI images can also be achieved within the RSG software.

Future work

Topics for future work include:

• The discussion document on SAR geocoding needs to be given high priority so that a choice of approach can be based on clear criteria.

• Ascending/descending SAR image pairs of BASUK and BASSW need to be obtained as soon as possible. This is to allow geocoding and information extraction flow-lines to be set up and tested well in advance of the 1998 melt season. Reference images also need to be selected and geocoded well before the melt season.

Develop a PCI EASI /PACE routine for registration of SAR images.

Figure 3.1: Flowchart of SAR snow mapping procedure.

Figure 3.2: Flow chart for surface classification of Landsat TM images (from Nagler 1996).

• WP 330: Remote Sensing Methods for Medium Resolution Sensors

Responsible: M. Baumgartner, UBE, months 2-12

Objectives: This WP will review, improve and document semi-automatic and automatic procedures for estimating albedo, and generating snow cover maps from data of medium resolution sensors; this study will include geocoding issues. The output will be in a format suitable for integrating into the general flowline concept of HYDALP.

WP332 and WP333 have not yet started.

Tasks within WP 330

WP 331 Review & improvement of methods for MROI UBE Baumgartner

WP 332 Water balance and evapotranspiration MLURI

WP 333 Documentation of methods UBE Baumgartner

The only active Task is WP 331- Review and improvement of methods for MROI.

• WP 331: Review and improvement of methods for MROI

Responsible: Michael Baumgartner, UBE, months 2-10

The review of the most commonly used methods for medium resolution sensors, especially NOAA-AVHRR data is ongoing. An extended summary will be presented in the next 3-month report. Within the next few weeks, the development of new approaches of using medium resolution remote sensing data, including geocoding and classification procedures, will start.

From visit to UBE: Ron Caves, SCEOS gave details, below.

R. Caves visited UBE on 25th and 26th June 1997 to get better acquainted with their work on MROI and its role in the HydAlp Project. The main steps in AVHRR processing were discussed.

Encoding:

AVHRR data is received as 10-bit and has to be stored as 8 or 16 bit. 8 bit storage involves an element of information loss. 10 to 8 bit conversion is achieved either by dividing by four or by scaling between minimum and maximum values. 16 bit storage involves no information loss but twice the amount of space is needed.

Calibration:

Radiometric and atmospheric correction are not needed for snow cover mapping as the snow albedo is generally much higher than that for other features. Calibration would be needed if snow quality were an issue.

Geometric correction:

AVHRR has a wide field of view and the distortions caused by the Earth’s curvature must be corrected. Distortions are so severe along the edges of the swath that it is advisable to use data only from the central third of the swath. Geometric correction can be neglected for small sub-images extracted from the centre of the swath. The only reference on geometric correction of AVHRR data known to MB is in German (Urs Frei, University of Zurich Remote Sensing Series). The availability of a reference in English needs to be investigated.

Geocoding:

Geocoding is the main processing problem for AVHRR. Geocoding must be accurate for time series analysis. The NOAA platforms, (there are several) are of an old design and do not have particularly stable orbits. This results in a marked variation in the equatorial crossing of each satellite. Hence, each AVHRR image has to be separately geocoded.

The equatorial crossing and other orbit information is listed in image headers. Geocoding can be based on these orbit parameters and 2-3 GCPs. However, this method has been found to produce misregistration errors of 5-6 pixels (i.e. 5-6 km), which is not sufficiently accurate for HydAlp. The PCI AVHRR Orbital Navigation routines are of this type. The geocoding approach adopted at GIUB is based on matching a much larger number of GCPs (15-20 per sub-scene). Where possible GCPs are selected from a reference set identified from previous images in a time series, though this is not always possible. This approach is time consuming particularly in mountainous regions - geocoding throughput is one image per day.

Classification:

Classification is based on a supervised learning routine. If the image is cloud free the visible and NIR channels are sufficient for classification. In the presence of clouds NIR and FIR are used. Problems arise when clouds and snow have similar temperatures, e.g. thin cirrus and around the edges of clouds (mixed pixels); the latter can cause halo like effects.

It is advisable to classify an image prior to resampling it, as nearest neighbour then becomes the automatic choice of method for resampling.

Improvements

Several planned improvements to existing methods were discussed. These included:

• Developing platform independent solutions - using RESURS data in addition to AVHRR.

• Further automating processing procedures for geocoding and classification..

• Utilising DEM data. Previous SRM simulations have just been based on elevation zones.

• MB stated that the routines currently used at UBE for AVHRR processing are not portable. This will cause problems if MROI methods are to be applied to BASSW or BASUK. MB is considering setting porting of the routines as a project for a research student.

• WP 400 Hydrological Modelling

Responsible: R.Ferguson, SCEOS, months 1-17

This work package has the following objectives:

• To review existing modelling methodologies and identify possible areas of improvement.

• To modify existing models to make optimum use of remotely sensed and ground information for calculating daily runoff using as input in situ measurements of hydrological and meteorological variables, information derived from remote sensing data, and - in the case of runoff forecasts - also meteorological forecasts. The work will rely on the runoff models HBV and SRM.

The work is broken down into the following sub-work packages:

WP 410 Intercomparison of the SRM and HBV runoff models.

WP 420 Compilation of runoff parameters.

WP 430 Methods for remote sensing and hydro-meteorological data fusion.

WP 440 Hydrological model modification.

WP 410 was the only work package operational in months 3-6. The work carried out is summarised below.

• WP 401: WP Supervision and Quality Control

Responsible: O. Turpin, SCEOS, months 1-17

WP supervision and quality control involves resolving any problems that arise and ensuring that the objectives of all tasks are met and that all deliverables are on time. The only other task currently operational, WP 411, is on time, with no problems thus far.

• WP 410: Intercomparison SRM/HBV Model

Responsible: O.Turpin, SCEOS, months 1-9

This work package has the following objectives:

• To review the conceptual and physical basis of each model, the parameterisations of processes, and the way the models use or could use remotely sensed information.

• To point out strength and weaknesses of the models and to make recommendations for model modifications.

• WP 411: Intercomparison

This task is not continuous throughout months 1-9. Some preliminary work was done in months 1-3. The main work has commenced, though only intermittently in months 5 and 6 because most of the project participants concerned have been on leave at one time or another. Points which are expected to be covered in the intercomparison were listed in the month 1-3 progress report and the list is not repeated here.

New work performed has included the following.

• The review of the current versions of the models (SRM: 1994, HBV: 1996) has continued. This is being done on the basis of the user manuals and executable code provided by UBE and SMHI. The user manuals are detailed and informative so it has not yet been necessary to investigate the source code or contact the programmers.

• A review is ongoing of research and development work done in several countries during the development of past and present versions of the models. This is being performed by reading journal and conference papers identified from the literature search done in months 1-3, and following up citations of other relevant publications. Some of these papers explain the basis of choices made in the development of SRM and HBV, and identify circumstances in which model performance is relatively weak. Other papers report experimental research altering particular aspects of a model. In some cases this led to changes in later official versions of a model, in other cases it did not but may still be relevant for future changes. Consideration is being given to whether previous results might be site-specific and not transferable to the full range of environmental conditions represented in HYDALP.

• A start has been made on the first version of the intercomparison document, on the basis of the literature reviews referred to above. This version will be distributed by email to the person or persons concerned with hydrological modelling in each partner institution (SMHI: BJ; UBE: MB and HK; IMGI: GG and TN). Comments will be requested to help us in SCEOS prepare the final version of the internal report due to be delivered in month 9. The draft document may also be useful to partners in other institutions who at present have experience of only one of the two models; the draft intercomparison will explain the quite considerable differences between the models.

• A start has been made on a review of other models for snow and ice melt. The immediate aim of this review is to relate HBV and SRM to a wider context and identify any respects in which either or both of them is unusual or novel. In the longer term, this review will be useful also for considering the fusion of remote sensing with hydrological modelling (WP 430) and possible changes to models (WP440). For this longer-term aim the literature review will be extended to both empirical and conceptual modelling approaches to the spatial distribution of snow water equivalent in snowpacks.

• The literature review for WP410 not only covers model structure and the use of remote sensing but also the philosophy and techniques of calibration, verification, testing, and sensitivity analysis. It was originally planned to leave this till it is needed later in HYDALP (WPs 420, 500 and beyond), but much of the literature about the models and changes to them also discusses calibration and testing so it seems sensible to assimilate this information now.

• The use of SRM and HBV was demonstrated at the Scottish Customer Focus Group (SGFG) meeting that was held in and near Aberdeen in late July as part of WP210. Useful feedback was obtained from the SCFG about the main customer requirements in BASUK, which appear to be similar to those already known for other basins (i.e., the main requirement is runoff forecasts). There is some interest also in short-term forecasts of changes in the snowpack. Local experts feel that sensible heat transfer is important in BASUK and that it may be advantageous to include windspeed in melt calculations, at least at sub-basin scale.

• Liaison has continued with UBE, SMHI and IMGI regarding data requirements of models and the implications for database management (WP201/220). This was the topic discussed at greatest length during the technical meeting in Aberdeen immediately after the SCFG meeting.

• It has been agreed with all researchers concerned with modelling tasks now or later in HYDALP that it will be useful to hold a technical workshop of modellers to discuss a range of issues. Fixing a time and place for this meeting has not been easy because of the absences of key people at various times in the summer and autumn either for leave or for research visits. The original plan of a meeting in month 6 or 7 has proved impossible and it now seems likely there will be a modellers meeting at the end of the general meeting in Bern in month 9. If a second meeting is felt necessary it is likely to be held at SMHI, probably in month 10 or 11.

• The review will also consider the use that has been made in previous studies of near-real-time updating to help improve the simulation or forecasting of runoff. Such updating may be through comparison of simulated with observed runoff, or comparison of modelled snow cover with that observed by remote sensing. In the past SRM has generally used a few to several snow cover images per season, but HBV has generally used none. A review of the temporal frequency of snow cover observation in previous work will provide a basis for later discussion (WP430) of the possibility of using more frequent EO input.

All these lines of work are ongoing.

A list of the parameters required to run the two models being compared, and the temporal resolution of these data are given as:

• WP 500 Model Validation and Application

Responsible: H.Rott, IMGI, Months 12-28

This work package has not yet been actioned.

• WP 600 Contributions to CEO Enabling Services

Responsible David Miller: MLURI, months 22-30

Objectives: Information provision: Centre for Earth Observation Demonstrator

Development of the Hydalp World Wide Web (WWW) site has been in two main areas. These are presentation and administration and documenting geographically referenced data.

1. Presentation and administrative materials within the WWW pages have increased to include minutes for the Scottish Customer Focus Group; a calendar which Participants can edit for a centrally available reference site; and current reports from the Hydalp project. Some illustrations of current datasets are also available. The results of the survey of Participant computer resources are available as are the reporting trees for the months of June and July 1997.

2. Development of the meta-data and meta-model pages is progressing with reference to the proposals made by EUROGI. The content of the pages is now believed to be appropriate for the purposes of Hydalp. An example page is included in the Appendices of the report for WP310. Design of the pages has been with a view to providing potential future users an informed impression of the nature, content and quality of both the data and the models being used (or developed) within Hydalp. By necessity, such pages have not yet been exposed to a wider audience and thus refinement for a wider applicability may be required.

However, several pages have now been completed for datasets within BASAT, BASCH and BASUK. These are currently at the stage of requiring to be checked by the Participants before release to the rest of the project.

The meta-model pages have not yet been used. Both of the hydrological models being used within Hydalp will be included in due course.

A listserver facility has been introduced for aiding electronic communications between participants, which is available as an alternative to individual mailings. Access to the listserver is only for Participants and requests for inclusion on the "list" should be sent to the , "list moderator" (a.law@macaulay.ac.uk) who is located at MLURI.

A file transfer protocol (ftp) site has been established at MLURI for the transfer of large digital datasets (address: highgate.mluri.sari.ac.uk; login: anonymous; password: anonymous; directory: hydalp). The advantage of such a site being based at the Participant's host computer is principally one of the ease of data management. Further use of this site is envisaged, at least for participants involved in WP310.

Current developments are focusing on the potential role of the HYDALP WWW site within the CEO ESWE site. Specifically, what links will be required between Hydalp WWW pages and those of ESWE for active demonstrations. This will be an important step because it will establish certain 'ground rules' as to how the group envisage the demonstrator to be used by outside parties.

All WWW that are provided by the Hydalp group are now linked into the project home page, at

http://www.macaulay.ac.uk/hydalp/

• Appendix

• Example of Metadata report (WP310)