IV. Plenary Session

Introduction

Following lengthy discussions on the following topics, it was decided that the GCOS/GTOS Plan for Terrestrial Climate-related Observations (GCOS-21) should be revised to reflect the discussions. Each discussion was led by a different Panel member who provided the Panel with a working paper on the subject. Following the discussions, small working groups prepared a report summarizing the presentations and discussions. The following are those reports.

Tier Sampling Scheme

The GCOS JSTC requested that the TOPC carefully review the tier concept and in particular re-evaluate the need for tier IV. Dr Robert Scholes prepared an evaluation of the system and led the discussion. At this meeting, the tier system was still based primarily on ecosystem characteristics. Subsequent meetings of working groups on hydrosphere and cryosphere are expected to further clarify issues. The following is a summary of that Panel’s discussion.

Rationale for the hierarchical system

The detailed description and rationale for the hierarchical system is given in GCOS-21 and the GTOS Planning Group Report. Briefly, it is not efficient to measure all variables at the same time interval, spatial scale or with the same accuracy. Since these three factors tend to be correlated, as a result of the characteristic space-time domains of key processes, an efficient sampling system naturally forms a hierarchy, with some variables measured continuously and intensively, but at a few locations, while others are measured simply, but at a large number of locations. When this continuum is matched against the types of observing systems which exist, the latter can be classified into five broad levels, identified as tiers. Some of the characteristic features of each tier are given in Table 1.

Table 1. Characteristic Features of Each Tier.

Implementation

The tier structure is a classification system to aid implementation, not a rigid formula for implementation. All the tiers are necessary, but not all the variables are represented at all tiers. For example, the hydrosphere has few variables at tier IV, because the characteristic time scale of hydrological processes tends to be too fast to benefit from infrequent sampling. The key tiers for early implementation are II, III and V, because they are served by existing structures.

The tier IV issue

Tier IV was designed to consist of in the order of 10 000 land surface points, which would be visited regularly, but infrequently for the collection of surface parameter data. It was intended to bridge the gap between satellite observations, which are quasi-continuous in space and time (tier V), and the 500 or permanently-staffed but non-optimally distributed ground stations (tier III). The main problem with tier IV is that it is new, and cannot be assembled from pre-existing systems. This makes it apparently expensive and technically untested. The question addressed by the Panel at this meeting was: can tier IV be dispensed with, greatly reduced, or phased in without undermining the validity of the entire plan?

The unique role of tier IV

Tier IV has the unique function of providing accurate and spatially-resolved data on variables which at present cannot be remotely-sensed. An example is the soil carbon content, which is important because it is the largest biospheric carbon pool, is subject to change, and influences other factors such as the water-holding capacity. It cannot be observed from space. Tier IV serves two primary purposes, only the second of which is strictly speaking a monitoring function, but both of which are necessary for an operational system:

a) One-time measurements, for purposes assessing the state of the system and model parameterization; and

b) Repeated measurements, for purposes of change detection.

Tier IV can also act as the calibration and validation points for indirect remotely-sensed variables. Tier III can also provide these data, but may have insufficient sites to calibrate or validate all the ecosystem classes. It is important to separate the unique roles from the additional roles, because they have different sampling requirements.

Sampling pattern

One issue regarding tier IV is whether the location of sites should be systematic (e.g., gridded), stratified random, or targets of opportunity. Each scheme has implications in terms of cost, political feasibility and statistical analysis. Table 2 provides an analysis of the financial, political and statistical sensitivities of the issue.

Table 2. Implications of Different Sampling Patterns.

Tier IV need not have a single sample scheme for all variables and all places. For instance, where the purpose of the observation is calibration and validation of a remotely-sensed variable, the target-of-opportunity approach is acceptable. For statistical change detection of a variable not indirectly measurable, the scheme must be stratified or systematic. A systematic scheme in one country remains compatible with a stratified scheme in another, if they are designed to the same accuracy specifications.

What cannot be done without tier IV?

The following variables are unlikely to be available with useful accuracy and resolution in the foreseeable future without tier IV:

Necromass

Soil carbon

Soil total nitrogen

Soil phosphorus

Soil texture

Rooting depth

Ground water storage fluxes.

The accuracy of the following variables, which are indirectly measurable or potentially so, will be seriously compromised by an absence of tier IV:

Biomass - above ground

Biomass - below ground

Roughness - surface

Vegetation structure

Land use

Soil bulk density

Soil surface state

Precipitation

Ice sheet mass balance

Permafrost - active layer

Permafrost - thermal state.

GTOS also has several non-climatic requirements from tier IV. These include:

a) Land use inputs

b) Disturbance regime

c) Soil chemistry (pH, nitrate, phosphate, bases, acidity, CEC).

Options for implementation of tier IV

While there are costs associated with analyzing and storing any collected data, reducing the number of variables in tier IV is not a useful strategy, since the costs involved in sampling have largely to do with accessing the sites, not the time spent at an individual site. There may even be a case for increasing the at-site data collection in order to be able to share the sampling effort with a wider range of clients. For example, could GTOS geo-referenced socio-economic data be collected this way? The main cost-reducing options are:

d) Fewer points;

e) Use targets of opportunity; and

f) Phase the implementation gradually.

i) Fewer points

There are two issues: accuracy and spatial resolution. The debate on tier IV thus far has focused on the number of sites needed to validate a land cover product with a given number of classes, to a given level of precision and confidence for each class. This is a relatively trivial exercise in binomial probability. For instance, to validate the 17-class IGBP land cover scheme to within + 15% accuracy with 95% confidence would require 425 points. The accuracy requirements and number of classes for TOP purposes are likely to be more stringent. A + 5% accuracy would need approximately 4 000 points, and a + 0.5% accuracy would need about 22 500 points.

Most of the variables which depend on tier IV are continuous values, not categories. The determination of an adequate sample size for these is more complicated, since it requires a knowledge of their statistical distribution, which is largely lacking, and varies from variable to variable. For a completely unbiased and efficient sampling scheme (stratified random, for instance), and a normally-distributed variable with a coefficient of variation of 30%, an accuracy of + 10% with 95% confidence would need 36 samples, + 5 % would need 144 samples and + 0.5 % would need 14 400 samples. In practice, most of the variables in question are log normally distributed, and then only once they have been stratified, so the sample number needs to be multiplied by the number of strata if each is to meet the accuracy criteria, or by some area- or value-weighted number for a given global accuracy.

If only a global estimate is needed, the sample number is greatly reduced; perhaps by 75%. If regionalized estimates are needed, the requirement goes up in rough proportion to the number of regions.

Without doing a rigorous analysis, it seems that the original estimate of the order of 5 000 - 10 000 sites for tier IV is still valid. However, where the tier IV sites are simply required to calibrate or validate a remote-sensing algorithm, the required sample numbers are much smaller, and the sample location requirements are much less rigorous. Typically greater than 30 samples each are required for calibration and validation of a continuous, linear model if the errors are normally distributed, the model is reasonably predictive (accounting for >75% of the variance), and the sample points cover the full range of variation.

If the number of points needed can be reduced to 500 - 1 000, then tier IV can be substituted by tier III, but all the problems associated with a biased sample scheme remain.

ii) Targets of opportunity

By piggy-backing on other activities, the costs of sampling can in theory be reduced to the marginal costs of the additional effort needed to collect the TOP data. An example is the Soil and Terrain Data (SOTER) project to improve global soil data. If the vegetation component were slightly enhanced, and geo-location specifications were tightened to + 10 m, many TOPC requirements would be met.

There are two main drawbacks with using targets of opportunity: the sample locations are likely to be biased, biasing the values in an unknown way; and the chances of being able to revisit the point at the same low cost are small. This approach could be useful for calibration and validation of indirect algorithms and for one-time parameterizations, but is not suitable for change detection except in the sense of archiving a current state, which future generations may find useful. At a minimum, the TOPC data system should make provision for the recording of tier IV-type data from activities outside of GCOS/GTOS, and should actively pursue their acquisition from the original collectors.

iii) Phased implementation

All of the tier IV variables have relatively slow rates of change, which allows them to be infrequently collected. Thus, only 10-20% of the target sample needs to be collected in a given year. By concentrating the data collection in a given year regionally, logistic costs can be reduced. If it is assumed that on average one point can be collected per day by two observers, one of which is provided by the host country, a single GCOS/GTOS employee could collect about 200 points per year; five employees regionally deployed could cover the world at the desired density. Alternatively, or additionally, each tier III station could be tasked with collecting one or two tier IV data points per year, in an a priori determined location.

Recommendations

If tier IV is not implemented, key variables which are directly or indirectly involved in the global climate system cannot be accurately collected. Most of these variables relate to below-ground processes with a high spatial variability and no obvious way of collecting via remote sensing. The sampling intensity cannot be greatly reduced without making the tier irrelevant. The real cost of tier IV is probably not as high as is thought, given the cost of activities which currently fill this role (such as satellite observations and their ground validation). The benefits of enriching the land surface data go beyond the needs of TOPC, or of GCOS and GTOS, into issues such as natural resource management at a regional or local scale. Tier IV can deliver national-level information and involvement. A specific action plan should:

1. Publish a brochure explaining and publicizing the Global Hierarchical Observation System for Terrestrial Ecosystems (GHOSTE);

2. Publish and distribute a methods handbook for tier IV data and actively encourage the placement of target-of-opportunity tier IV data in public domain databases; and

3. Phase in the implementation of tier IV by making it a tier III responsibility. Appoint one or two dedicated tier IV observers, who will simultaneously do in-field training of tier III personnel and collect tier IV data.

Guidelines for Participation of Sites

Prof. Shidong Zhao led a discussion of the requirements for participation of sites in the GCOS/GTOS network. He first presented the Chinese system and their requirements and made several suggestions for GCOS/GTOS. The following is a summary of the discussion.

Issues

All GCOS/GTOS sites have to satisfy the conditions of representativeness and sufficiency. Tier I sites are large experimental areas and do not fall within the scope of a long-term monitoring project. The responsibility for these sites rests solely with research programmes such as the International Geosphere-Biosphere Programme (IGBP), Global Energy and Water Cycle Experiment (GEWEX), etc. The responsibilities of GCOS and GTOS should be to assure that the data are available and that there are links to the data. Further lessons learned at these sites should be incorporated into monitoring programmes, where appropriate.

Tier II sites should be located near the centre of the range of environmental conditions (though not necessarily near the centre of the geographical range) of the system which they are representing. There are never likely to be so many tier II sites that choosing between them is likely to be a major problem.

Tier III sites are intended to sample the range of variation present in the system which they represent. This means that some of them will be close to the average of the various environmental factors which make up the environmental range of the system, while others will be closer to the extremes, and perhaps even at the ecotone of transition to a different system. There are very many potential tier III sites, but they are not optimally distributed. As a result, some ecosystem types may have more potential tier III sites than are needed for GCOS/GTOS purposes, particularly if a rough global balance is to be preserved. Other ecosystem types may have too few sites, or none at all, and thus GCOS/GTOS will need to work with funding organizations to enhance the network.

GCOS/GTOS must simultaneously have a top-down and bottom-up approach to site selection. The bottom-up approach is implemented by publicizing the existence and benefits of GCOS/GTOS, and inviting sites (and pre-existing networks of sites) to apply for membership. The GTOS Secretariat, acting on selection criteria defined by the Panel, would then screen the applicants for suitability. The criteria are spelled out in the GTOS Planning Group Report. Briefly, they are:

Each participating country should have at least one per biome;

They should be capable of collecting the appropriate data;

Sites in under-represented systems have priority over already-represented systems; and

Reasonable permanence is required.

All else being equal:

g) Sites where research is also carried out are preferred;

h) Long-established sites are preferred;

i) Existing sites are preferred over sites which need to be established;

j) National support for the site is preferred to sites dependent on external funding; and

k) Accessible, practical sites are preferred.

The bottom-up approach would not fill gaps where no or insufficient sites exist, or where the sites are not aware of the existence of GCOS/GTOS. A top-down approach, is needed to fill gaps in ecosystem types where no sites have applied for membership. This will first require a search of the Terrestrial Ecosystem Monitoring Sites (TEMS) (http://www.wsl.ch/services/services.html) database to determine site location. The GTOS Secretariat will then need to contact these sites asking if they would be willing to participate. If there are insufficient candidates in the database to sample some ecosystems adequately, then an active process needs to be followed to upgrade or establish sites in that type. ‘Adequacy’ is obviously relative - an initial target should be 10-30 tier III sites per system type. ‘System types’ are initially defined as:

l) The 11 IGBP natural land cover types for terrestrial systems (evergreen needleleaf forests, evergreen broadleafed forests, deciduous needleleaf forests, deciduous broadleafed forests, mixed trees and shrubs, closed shrublands, open shrublands, woody savannahs, savannahs, grasslands and permanent wetlands);

m) The most important foods, fibre crops and agricultural systems (rice, wheat, maize, potato, sorghum and millet, cassava, sugar cane, extensively-grazed livestock, vegetables, tropical fruits, temperate fruits, and cotton);

n) The aquatic systems are rivers, lakes and estuaries; and

o) The cryospheric systems are permafrost, ice sheets, ice caps and glaciers.

The site location of tier IV sites are based on an unbiased statistical representativeness. It is impractical to force a single statistical design on all countries. Hence, individual participating nations would be responsible for locating the sites, and may choose either a systematic or stratified-random approach (or both, for different variables or different systems). This requires an a priori location specification, but permits rejection of the site and resampling out of the same population if:

p) The site is inaccessible; and

q) Sampling at the site would compromize national interests.

Management

The general management structure outlined in the GTOS Planning Group Report is appropriate, but in its initial implementation it should focus on establishing an effective GTOS Secretariat and strong links to national implementing agencies and networks. The steps in an initial operating strategy for managing the system are:

1. Hold an international meeting of site and network managers. Its purpose would be to establish and ratify the rules and methods of data collection, data sharing and quality control;

2. Charge the GTOS Secretariat, in conjunction with discipline and system experts, to develop a “methods’ manual”, reporting procedures and a training programme. Training exercises should be in-field (not central), and should double as data-collection exercises; and

3. At the same time, the communication and distributed database functions of GCOS/GTOS need to be established so that they are ready to handle data flows.

Standardization and quality control

For each variable, standardization should be primarily achieved by specifying the necessary accuracy, frequency, and spatial footprint. Standard methods can be encouraged by publishing manuals and providing training, but are not obligatory. Full descriptions of the methods used at the site should be provided by the site to the GTOS Secretariat at the time that a site joins the network, and updated when necessary. Variables where there is a significant bias due to the method used should be subject to expert review and cross-calibration exercises.

The first and most important line of quality control is the point of data collection. Data items not accompanied by time, exact location (+ 100 m) and method should be rejected. On entering a GCOS/GTOS data system, data should be passed through a “smart” filter to detect gross errors and inconsistencies, and then be checked for reasonableness by a discipline expert.

Data management

The general principles of data management outlined in the GCOS and the GTOS Planning Group Report and their accompanying documents must provide the guide.

Briefly:

All data entered into the system are public domain information. Data providers should exercise their prior publication rights within a reasonable time (two years) before entering the data into the system;

Data users should be charged costs of filling the user request only;

The GCOS/GTOS data system will be distributed. The GTOS Secretariat should be responsible for maintaining meta data, pointers to other data sets, and data sets which have no other home; and

Special efforts must be made to accommodate the data access needs of countries which are not well-connected to the electronic network. The strategies should include: establishment of regional mirror sites with regional data sets; the placement of key data sets on CD-ROM; use of email, fax and postal services for the input and export of small quantities of data where necessary and feasible.

Recommendations

1. Begin a dialogue with existing site networks (either international or national) to produce a draft of workable data exchange rules and procedures and to refine the potential site database;

2. Make the TEMS database more comprehensive by asking regional and national experts to check and populate it;

3. Conduct an initial tier III selection according to the procedures described above;

4. Identify under-sampled regions and ecosystem types; and

5. Invite site and network managers of the identified sites, plus representatives of organizations which could establish or upgrade sites in under-sampled regions, to a conference to agree on their participation in a network.

Cryosphere Needs - the Next Steps

Dr Haeberli provided a review of the cryosphere section of GCOS-21. The following is a summary of the recommended updates to GCOS-21.

Issues

There is now high confidence that many components of the cryosphere react sensitively to changes in atmospheric temperature because of their thermal proximity to melting conditions. The varying extent of glaciers has often been used as an indicator of past global temperatures. In fact, obvious thinning, mass loss and retreat of mountain glaciers have taken place during the 20th century. The areal extent of the Northern Hemisphere continental snow cover has decreased since 1987 even though there is much variability from year to year, and no definitive long-term trends can be defined. Climate projections into the coming century indicate that there could be pronounced reductions in seasonal snow, permafrost and glaciers with a corresponding shift in landscape processes. Such reductions would have significant impacts on related ecosystems and socio-economic factors. The thickness of the active layer of permafrost could increase and extensive areas of discontinuous permafrost could disappear in both continental and mountain areas. More water would be released from regions with extensive glaciers, and both engineering and agricultural practices would need to adjust to changes in snow, ice and permafrost distributions.

Priorities

Priorities with respect to initial implementation of monitoring cryosphere variables are based on: 1) climate relevance, 2) feasibility of measurement, and 3) the existence of suitable sites and programmes. The variables listed below are described in GCOS-21.

Sea ice

Sea ice concentration

Sea ice motion

The variable sea ice concentration includes sea ice extent. Operational elements exist (SSMI data stream, Arctic buoy network) for initial implementation for both of these variables. Priority for initial implementation is mainly related to climate feed-back (albedo) and practical implications (shipping).

Snow

Snow cover area and snow water equivalent.

Operational structures exist (weekly product from AVHRR) for initial implementation. Priority for initial implementation is mainly related to climate feed-back, to intense interactions with all other systems (glacier mass balance, permafrost thermal state, surface water, soil humidity, etc.) and to practical implications (hydrology, tourism, etc.).

Ice sheets

Ice sheet geometry

Ice sheet surface balance

Ice sheet geometry (elevation, margins) combines the former variables (GCOS-21) ice sheet extent/topography and ice sheet mass balance, whereas ice sheet surface balance includes firm density profiles as mentioned in the GCOS Plan for Space-based Observations (GCOS-15). Ongoing activities include studies of possible errors in altimeter measurements to monitor mass balance, at least in those areas where the slope is not so steep that the altimeter loses focus. It will be important to measure ice sheet velocity at points distributed over the whole of a satellite image to ensure the most accurate calibration.

Glaciers and ice caps

This variable includes information on mass/volume, length and area. Glacier fluctuation is one of 27 internationally selected geo-indicators of rapid environmental change. Glacier shrinkage at decadal to century time scales is a worldwide phenomenon, influences the water cycle and natural hazards in cold mountain areas, reflects an additional energy flux roughly comparable to the estimated radiative forcing, and, hence, constitutes a key indicator of ongoing climate change. Measurements should continue on glacier mass balance as a direct climate change signal, glacier length change as a delayed but also enhanced and more easily determined climate signal; and glacier spatial distribution patterns (glacier inventories) for assessing regional effects. Priority relates to the exceptionally clear and easily understood climate signal presented by glacier changes, to the practical impacts at local, regional and global scales. Problems mainly relate to reaching global coverage and to monitoring large glaciers especially with respect to sea level change.

Lake and river freeze-up and break-up (timing)

There are no recommended changes to GCOS-21 or updates regarding this variable. Priority for initial implementation is based on the characteristics of the variable as an integrated long-term signal that is easy to measure.

Permafrost

Permafrost active layer

Permafrost thermal state

Frozen ground activity and sub-surface temperature regimes are two out of 27 internationally selected geo-indicators of rapid environmental change. Efforts of the International Permafrost Association with respect to long-term monitoring concentrate on the rescue of borehole temperature data, on Circumpolar Active Layer Monitoring and on mountain permafrost monitoring. Standard International Tundra Experiment (ITEX) procedures recommend landscape-level observations at or near the end of the thaw season on a standard-size grid. Seven sites are presently being observed in the US Arctic, 7 in the Russian Arctic, 1 in Canada, 1 in Sweden and 1 in Norway. Inter-annual comparisons of thaw from within a site and among different sites within a region show significant season-to-season variations. As more sites from different programmes are added to the CALM network and longer-term records are accumulated, the significance of climate fluctuations on regional patterns of thaw and on plant and soil relationships and processes should be better understood.

Argentina, Canada, Norway, France, Germany, Switzerland, Austria, Italy, Kazakstan (Central Asia) and China are participating in an effort to monitor mountain permafrost. Data input to the Global Geocryological Database (GGD of IPA) is planned for 1997/98. Permafrost temperatures at about 15 m depth in the few presently existing boreholes appear to have increased in the European Alps, the Kazak Tien Shan, the Kirghize Tien Shan and in the Qinghai-Tibet area.

Priority for initial implementation is based on the importance of the feed-backs involved with changes in active layer depth (CH4 emission, soil moisture, growth conditions), practical applications (stability of foundations for roads, pipelines, buildings, etc.) and the existence of monitoring structures (CALM).

Sampling strategy

A sampling strategy for the cryosphere still has to be developed. Sea ice concentration and snow cover area would fit into tier V of the presently existing system; ice sheet surface balance, glaciers and ice caps (extent), permafrost active layer and permafrost thermal state at depths greater than about 15 m into tier IV; and permafrost thermal state at depths shallower than about 15 m, snow water equivalent and extensive glacier mass balance measurements into tier III. Uncertainties continue to exist, for example, with measurements of elevation of ice sheets and large glaciers or with the mapping of ice sheet margins. The corresponding questions could best be treated by a short meeting which could take place in connection with the Symposium of the International Glaciological Society on the Representation of the Cryosphere in Climate and Hydrological Models, Victoria, B.C., Canada, 12-15 August 1996. This meeting could profit from the results of a joint International Arctic Science Committee (IASC)/Scientific Committee on Antarctic Research (SCAR)/World Climate Research Programme (WCRP)/International Commission on Snow and Ice (ICSI) Workshop on Ice Sheets, Glaciers and Related Sea Level Change, Fjaerland, Norway, 21-22 June 1996. Cooperation between ICSI(IAHS) and IPA for developing a coordinated sampling strategy and for guiding the implementation of the present cryosphere observation plan could probably be established during such a meeting.

Recommendations

1. Continuation of existing monitoring programmes for snow, sea ice, glaciers and permafrost active layer;

2. Further development of monitoring programmes for ice sheets, permafrost thermal state and lake/river ice;

3. Coordination of an integrated cryosphere monitoring programme under the guidance of the International Commission on Snow and Ice (ICSI/IAHS) and the International Permafrost Association (IPA);

4. Assess the utility of the tier sampling concept for the cryosphere and develop an overall sampling strategy; and

5. Modify GCOS-21 to reflect the above discussions.

Precipitation

Issue

The JSTC of GCOS asked the TOPC to better clarify why there is a need for many kinds of time series of precipitation for terrestrial purposes, in contrast to relatively few that are needed to address atmospheric purposes. Also, the TOPC was asked to comment specifically on whether the precipitation data product proposed by the Global Precipitation Climatology Project, consisting of a data set of aggregated precipitation on a 1 d x 1 d spatial grid and at a 3-hour temporal resolution would meet the needs of the various terrestrial process communities. Dr Jurate Landwehr prepared the initial working paper and led the discussion.

Discussion

It is difficult to generically characterize what precipitation data are needed for hydrologic modelling/analysis regarding climate. The hydrologic cycle operates at many scales and levels of temporal and spatial resolution. Requirements for precipitation information are specific to the problem one is addressing. Macro-scale questions, such as delineation of continental or large regional water balances, may indeed find the proposed 2.5 d by 2.5 d product to be sufficient. But if one moves to a meso- or micro-scale, for example, to do catchment or watershed modelling, one needs to characterize the small scale variation of precipitation over the watershed. This calls for time series with higher temporal and spatial resolution which can only be obtained with in situ measurements aligned with watershed or catchment boundaries.

Furthermore, in situ measurements are necessary to provide a check on any satellite-derived precipitation product, as well as forming the basis for aggregated regional climate products such as the National Oceanic and Atmospheric Administration (NOAA)/National Climatic Data Center (NCDC)’s Historical Climatology Network (HCN) for temperature and precipitation data aggregated on the basis of the USA’s climatological divisions.

Dr Landwehr proposed a three-level structure (macro-, meso-, and micro-scales) for the exchange of global precipitation observations for the purposes of hydrological modelling directed to questions of climate variation:

For macro-level analyses, the future 1 d by 1 d GPCP product might be sufficient initially, although the 0.5 d x 0.5 d resolution being called for by SVAT modellers would be more flexible in fitting to hydrological drainages and route modelling; and

For meso- and micro-scale analysis, develop initially a set of as many long-term stations as available that are of good (measurement) quality and preferably hourly resolution (although daily and even monthly might be accepted).

The spatial resolution requirements are difficult to generalize and need to be set depending on the terrain - they could range from 1 per sq. km in a mountainous areas with small watersheds to 1 per 104 sq. km in terrain with low or no orographic features and homogeneous soil and drainage characteristics that allow aggregation to larger watersheds. The stations should be organized or identified with respect to watersheds or water-divides, and a single major climate zone.

Recommendations

GCOS/GTOS/TOPC would have to provide some minimal criteria for what constitutes a “good quality precipitation station”;

2. For all other purposes, maintain a meta data with pointers:

(i) to in-situ data centres that are holders of information that would be made available at the cost of filling the user request only; or

(ii) to specialized data sets already developed for climatic purposes, such as the HCN developed at NOAA/NCDC; and

3. The TOPC should further study the need for a.5 by.5 degree precipitation product, and if necessary and feasible, urge the GPCC to produce such a product rather than the 1 degree by 1 degree product that is planned.

Soil Moisture

The JSTC asked the TOPC to review the need for soil moisture data and whether or not it felt any soil moisture measurement was feasible to be observed from space. Dr Josef Cihlar and Mr Ken-ichi Kuma both prepared background documents. Mr Kuma led the discussion on the review of the need for soil moisture data, and Dr Cihlar led the discussion on the feasibility of observing soil moisture from satellites.

Issues

For climate prediction by fully coupled atmosphere-ocean-land models, sea surface temperature (SST) and soil moisture are the most important factors. The temporal changes of these two variables are large even for the short-term (seasonal) climate prediction so that their variation must be specified properly. As to the ocean, the current data assimilation techniques which have been implemented operationally, give the initial condition for coupled ocean-atmosphere models. These models can predict the evolution of the ocean and the atmospheric parameters in the climate time scale (from seasonal to decadal). Since SST is successfully observed by ships and satellites, the results of the model can be validated with these observed data.

A similar problem arises for the terrestrial environment and its role in climate prediction. With a coupled soil-atmosphere model, we need to predict the evolution of atmospheric as well as surface hydrological fields, e.g., soil moisture and run-off. Compared with the current ocean-atmosphere models, the development and validation of sound soil-atmosphere models has two problems to solve.

How to obtain the initial conditions for the soil moisture?

How to validate the soil moisture predicted by the model?

One of the difficulties in the first issue is that the initial condition must be well balanced with the coupled model. For this requirement, it is recommended that the initial conditions are obtained from the data assimilation using the same coupled model. To deal with the second problem, data sets obtained independently of the model are required. Thus, global soil moisture data sets are an essential requirement for the validation of the coupled models and thus successful climate prediction.

Apart from the climate model validation issue, soil moisture information is very important for hydrosphere and biosphere. It is a key determinant of primary productivity and vegetation structure, composition and density. From the hydrological view, soil moisture affects the partitioning of precipitation into soil moisture available for plant growth, ground water, and run-off.

Discussion

The soil moisture budget is determined by the balance of input (precipitation) and outputs (evapotranspiration, run-off, ground water storage). These components can be determined from the atmospheric forcing (radiation, temperature, moisture and precipitation) by means of a numerical soil model. Given proper atmospheric forcing to a sufficiently sophisticated soil model, one can obtain good estimates of soil moisture, run-off and the ground water storage terms. This simple soil moisture model-based method, which does not require the intensive computing (integration over time) of the coupled atmospheric model, has the advantage of low cost. A centre without massive computing power can apply this method to estimate soil moisture reserves and their temporal changes. On the other hand, soil moisture forced by the atmosphere may not be applicable to the initial condition for the soil-atmosphere model, where two-way interaction between soil moisture and atmosphere is important.

For the climate prediction by the coupled soil-atmosphere model, time integration should start with the soil moisture derived in the 4-dimensional data assimilation (4DDA). In 4DDA, both in-situ and space observations are integrated into global gridded data using a numerical model. Developed initially for numerical weather prediction, the 4DDA analysis is now regarded to be a very powerful way for climate monitoring. The feasibility of estimating soil moisture in 4DDA has been demonstrated.

Since soil moisture can vary rapidly in both space (metres) and time scales (hours, days), its direct or continuous measurement is not practically feasible at the global level. This has prompted interest in the development of remote sensing methods based on satellite sensors. Prior to this TOPC meeting, a survey of experts from several countries has been carried out to establish consensus on the state-of-the-art and prospects of remote sensing of soil moisture. Results of the survey are given in Annex 5 and can be briefly summarized as follows:

Remote sensing methods are capable of providing estimates of soil moisture for the near-surface layer only (5 cm or so), provided that vegetation is absent or the above-ground biomass is low. This would limit acquisition of useful data to arid, semi-arid and cultivated regions and exclude most forested regions;

The most accurate and robust technique at present is based on low-frequency passive microwave radiation measurements. Optimum sensors for soil moisture remote sensing have been proposed but not approved for satellite missions.

Since the remote sensing approach represents the only viable option for collecting global data sets, its thorough evaluation is necessary. This could include the possible use of soil moisture measurements from lower tiers as part of a combined validation/monitoring approach.

Recommendations

1. Establish the usefulness of near-surface soil information for climate monitoring purposes and the required spatial resolution should be established (as absolute quantities or in relative units). This implies the need for 4DDA tests to determine how the near-surface information would be used as initial or boundary condition in the models, assuming that it is available daily or every few days;

2. Based on the requirements, undertake a comprehensive evaluation of an end-to- end system (raw data to global data sets), and identify critical elements; and

3. Employ coordinated national initiatives to pursue further work on the critical elements.

Biomass Data Set

The need for biomass data was discussed at the meeting. Dr Rik Leemans led the discussion. There was considerable discussion over the need for biomass data. The topic needed further development so a small working group has been tasked with reviewing and rewriting this section of GCOS-21.

Requirements for Biogeochemical Cycle Models

It was noted that the biogeochemical cycling section of GCOS-21 was inadequate. Dr Robert Scholes prepared a discussion paper on the topic and led the discussion. The following is a summary of the discussion.

Why is biogeochemical data important to the climate?

Biogeochemistry studies the cycling of the elements essential to life. These cycles are strongly influenced by living organisms, and by interactions between the biosphere, atmosphere, hydrosphere and geosphere. Biogeochemistry focuses on the carbon nitrogen, phosphorus and sulphur cycles, and underlies all primary production and trace gas emission models. The impact on the climate is via the emissions of radiatively-active trace gases which occur as a result of biogeochemical cycling (CO2, CH4, N2O, ozone precursors and aerosols) and via the land surface characteristics such as biomass and leaf area which are constrained by biogeochemical considerations. Since biogeochemical cycling is strongly influenced by climate, this constitutes one of the major avenues for both impacts and feedbacks.

To understand, model or monitor any cycle requires three types of information:

r) The size of the pools;

s) The magnitude of the fluxes; and

t) The factors which control the fluxes.

In practice, all cycles can be simplified to some degree by considering only the major fluxes and pools. Generally, one flux will be ‘rate-limiting’, and its controlling factors need to be known in detail. For example, carbon assimilation by plants is a rate-limiting step in the carbon cycle, while nitrogen mineralization is limiting in the nitrogen cycle.

The crucial importance of the carbon cycle is well established, but why are the other elements included?

N Nitrogen is the principal limitation for the operation of the carbon cycle in most terrestrial ecosystems, including agroecosystems. N2O and NO are climatically-important trace gases.

P Phosphorus is the principal limitation for the operation of the carbon cycle in many tropical terrestrial ecosystems and most aquatic ecosystems. Phosphorus also controls the input of nitrogen to ecosystems via biological nitrogen fixation.

S Sulphur cycles in tandem with nitrogen and carbon. Sulphur aerosols are very important to the radiative balance of the earth and are the main component of acid rain.

Table 3 contains a list of the key variables needed for biogeochemical modelling and monitoring. Only the principal pools and fluxes have been considered. What follows is a list of the key variables that are needed in addition to those already listed in GCOS-21.

1. Plant tissue N [and P and S if possible] content, (mg/kg) and specific leaf area (m²/kg) on recent leaf litterfall, once a decade on the dominant species at tier III terrestrial sites. These are crucial variables for the modelling of NPP via biogeochemical models, and are subject to change under an elevated CO2 world. Accuracy + 5%.

2. Maximum stomatal conductance (mmol/s/m²), carbon assimilation rate (mmol/s/m²) and dark respiration rate (mmol/s/m²) of recently fully-expanded leaves of the dominant species, once a decade at terrestrial tier II sites. Irradiance, leaf temperature and vapour pressure deficit and nitrogen content to be specified at the time of measurement. These are crucial variables for both carbon models and hydrological models, and are subject to change in an elevated CO2 world. Accuracy + 5%.

3. Total N and Total S content of rainfall (mol/litre), accumulated daily and analysed monthly at tier II terrestrial sites. These are the major constituents of ‘acid rain’, which can be of either anthropogenic or biospheric origin, and constitutes a major input to the N and S cycles of ecosystems. Accuracy + 5%.

4. Tropospheric ozone is needed to validate the emission fields of ozone-forming gases. Daily data at a resolution of 50 x 50 km are desired, with an accuracy of + 20%.

5. Emissions of sulphur and other aerosols from volcanic sources and the height at which they are injected to the atmosphere, by major location and month. This may be achieved by a pointer to the appropriate data set.

6. Anthropogenic emissions of CO2, CO, CH4, NOX, N2O, NMVOC and SO2 every 5 years per nation. These data (except SO2) are reported under the Framework Convention on Climate Change. SO2 is widely reported by the major emitting countries in the context of transboundary pollution, and can be inferred from energy statistics in non-reporting countries.

7. Use of N and P containing fertilizer, by year, by nation or district within nation. These data are reflected in FAO statistics.

Table 3: List of Key Variables Needed for Biogeochemical Modeling and Monitoring

Prioritization of Variables

The Chairman, Dr Josef Cihlar led a wide-ranging discussion on the best way to determine the highest priority variables for implementation. Appendix IV contains a complete list of variables as prioritized by the TOPC. The criteria that were finally selected for prioritizing the variables were:

• What is the impact of the variable in both depth and breadth?

• How well will the variable help meet a specific GCOS objective?

• What is the cost of implementing? and

• What is the existence of implementing structures?

Table 4 lists those variables that at least one group ranked with both Priority 1, to meet a GCOS objective, and Feasibility 1. There were no variables that all three groups ranked with Priority 1. There are several, however, e.g., snow cover area, that were ranked 1 by two groups and 2 by the third.

Table 4. Short List of Priority Variables.

Key to Table 4:

Priority: P1 = high; P2 = medium; P3 = low.

Use: a = climate detection; b = impact assessment; c = prediction and simulation; d = model validation.

Feasibility: F1 = high; F2 = medium; F3 = low.

Data and Information Management

Prof. John Townshend made a presentation on the GCOS Data and Information Management System and led a short discussion on TOPC data and information management needs. The following is a summary of his presentation and the subsequent discussion.

Issue

Version 1.0 of the GCOS Data and Information Management Plan has been completed. This Plan describes a template for GCOS data management activities in broad terms. The information management strategy is based on the use of existing facilities and expertise. The system will be developed, implemented and operated by existing national and international organizations and programmes. The system relies on distributed centres sponsored by nations and institutions participating in GCOS. Participating data and information centres will be encouraged to develop data servers which can interact with clients according to agreed-upon guidelines. When fully implemented, the GCOS data and information management system will allow participating nations and programmes access to a wide variety of climate-related data. The GTOS proposal describes a similar approach.

There are a number of issues that have been identified by Prof. Townshend for the TOPC to consider. These include:

• The terrestrial community in terms of data management is not well served for many observations. This community urgently needs the assistance of the Data and Information Management Panel to remedy these deficiencies. A related but separate problem is the ability of centres to acquire the data they need for the user communities and to have access to it in appropriate formats;

• It is important that the needs of countries with inadequate access to the Internet have their needs met by the provision of data in other formats;

• For the terrestrial community, there is a lack of adequate data handling capabilities especially within the ecological and hydrological areas. There are some nominated centres such as the GRDC and GPCC, but further capabilities are needed. A World Data Center (WDC) for Soils is in existence. UNEP’s GRID holds several data sets collated from other sources;

• For ecological observations, the situation is less satisfactory though there have been proposals within WDC-A to help rectify the situation;

• There is a continuing problem in data being made available to centres in the appropriate format, such as long-term records collected for synoptic purposes which are in analogue format; and

• It is important that the needs of countries with inadequate access to the Internet are satisfied by the provision of data in other formats. While Internet access is increasingly available, problems associated with bandwidth and charges are likely to inhibit its use by many users for several years.

Recommendations

1. TOPC should participate in the proposed IGBP-DIS/WDC workshop which will be evaluating the needs of the IGBP community for data management;

2. The TOPC, with the assistance of the Data and Information Management Panel (DIMP), should identify and assess existing data centres to determine if its needs can be met;

3. The possible benefits of archiving more of the data collected through the World Weather Watch GTS system should be explored, especially in view of the development and application of web technology; and

4. Currently the land community does not appear to be well represented on the DIMP and this representation should be improved. Each member of the Panel should suggest names to the JPO for consideration as DIMP members.

Review of Space-based Observations

The JSTC had noted some apparent inconsistencies between the GCOS Plan for Space-based Observations (GCOS-15) and GCOS-21, and asked the TOPC to review their needs. Prof. John Townshend updated the Panel on GCOS interactions with CEOS and led a discussion of TOPC needs.

The GCOS-15 was presented to the Committee on Earth Observation Satellites (CEOS) in Montreal in October 1995. As a result of the presentation, CEOS has begun to rethink its role. The importance of the global observation systems is being substantially raised. The meeting decided to begin to develop an overall Integrated Global Observing Strategy (IGOS). The first meeting of IGOS is in Seattle (USA), during the week of 25 March, 1996. The results of this meeting will be reported to the next CEOS plenary in Canberra, Australia, in November 1996.

Recommendations

As a result of the discussions, the Panel had the following set of recommendations:

1. The various incoming and outgoing short- and long-wave observational needs should be reconciled with those of the AOP in terms of terminology and requirements, and then a consolidated list be presented to the Space-based Observation Panel;

2. The fire product should be redefined to include:

i) Coarse resolution global sensing of the distribution and daily frequency of fires;

ii) Fine resolution spatial calibration of fire scar extent using fine resolution data from satellites such as SPOT and Landsat; and

iii) Land cover data to provide information on the character of the fuel to assist estimation of the properties of the resultant atmospheric fire products;

3. Albedo should be added as a variable indicating how it is derived;

4. Emissivity is required by at least one set of users for interpretation of remotely sensed data but it is unclear whether this needs to be estimated by space techniques. TOPC should review and clarify this issue;

5. Surface roughness. Unclear currently whether this is required with a precision higher than that obtainable by parameterization using land cover classes;

6. Sea ice motion. This key variable is currently not in the Space Plan and should be added;

7. Surface water storage fluxes is not in the Space Plan but can benefit from space observation in its estimation (through observing areal extent of water bodies);

8. The observational requirements, nomenclature and definitions of ice sheet mass balance, ice sheet elevation and topography need to be clarified; and

9. Land use needs to be modified to better reflect the relative contributions of space and other observations.

4.12 The Demonstration Project

Dr Hal Kibby made a short presentation of a proposed demonstration project. The following is a summary of the discussion.

Issue

The issue is how to best demonstrate the need for GCOS and GTOS data in developing countries of the world. A project to demonstrate the utility of such data was discussed. The objectives are: (1) to demonstrate the operational feasibility of implementing GTOS and the terrestrial aspects of GCOS; (2) to demonstrate the need for GCOS, GTOS, and GOOS global data sets in seasonal to interannual climate predictions (in cooperation with WMO’s Climate Prediction Services (CLIPS) programme); (3) to demonstrate the utility of GTOS, GCOS and GOOS data sets in a national assessment.

Recommendations

1. The Panel recommended that in addition to spending time and effort developing coordination among potential GCOS/GTOS sites in Southeast Asia, some time should be spent on the development of tier III sites globally;

2. Prior to contacting the sites, clear guidelines for criteria for participating sites need to be provided; and

3. Identification and coordination of tier III sites should proceed simultaneously with development of the demonstration project in Southeast Asia.

Implementation Issues

As Chairman of the JSTC, Prof. Townshend felt that the implementation of in situ measurements presented a far grater challenge than the implementation of space-based observations. He pointed out that there is no overall coordinating mechanism analogous to CEOS for in situ measurements. Even if there were, the complexities are far greater because we are dealing with many more institutions.

It is important for the TOPC to take the responsibility of beginning to implement the in situ aspects of terrestrial observations. It is clear that TOPC needs to develop a statement that clarifies the vested interest that in situ sites have in participating in the programme. It was pointed out that one such interest would be in exchange of in situ data for satellite data. The satellite programmes need in situ data as much as the in situ programmes need satellite data. There are three potential options:

1. Where existing programmes are operating use these programmes;

2. Persuade an international agency to do it; and

3. Where all else fails invent a system ourselves.