The following is a collation of responses to the prototype review
conducted in 1992. It has been minimally edited to preserve the
anonymity of the reviewers. Personal communications and specific
editorial corrections for the documentation were also deleted,
however they were taken into account in the re-design and production
of subsequent versions of the database. The reviews are organized
according to questions on the response form that was distributed
with the first prototype, plus any additional comments provided
by the reviewer. The numbers refer to the reviewer, so that a
given reviewer's comments can be matched between questions (sometimes
a comment will make reference to answers in the previous questions).
I. WHAT ARE YOUR COMMENTS ON THE OVERALL USEFULNESS OF THE
DATABASE EFFORTS (E.G., IS THE DATABASE HELPING YOU IN YOUR RESEARCH
ACTIVITIES, IS THE PUBLISHING OF AN INTEGRATED GEOGRAPHIC DATABASE
USEFUL, ETC.)?
1) N/A
2) N/A
3) First, I want to congratulate you and your team with the compilation of such impressive data collection. My first feeling however, was a slight disappointment. Most databases included are databases that I already had checked, reviewed, and/or used in one stage or another of my research. Then I realized how much time it had taken me to locate datasets, read and reformat tapes, incorporate them into my own systems, present the data on maps and finally evaluate the data. This compilation is an enormous step forward for global (change) research. The current IDRISI format allows any user to look at the data in a glimpse, without much knowledge of programming, data structure etc. I think that this is the most important achievement of this data compilation. Although it is far from ideal, it will surely ameliorate the use of global datasets, lead to an improved data-use in global (change) research and eventually facilitate and promote the creation and development of better and more comprehensive global datasets.
I would like to see an exercise session included in the manual. Probably most users will not have much experience in GIS-technology and database techniques. A step-wise approach to guide the user to mapping, displaying and overlaying routines will promote a better understanding of the datasets and lead to more comprehensive applications. I have seen such a manual together with the IGBP-DIS [IGBP Data and Information System] pilot project "Global Change Database" and found it very useful. The preparation of such a manual should not be such large effort, because you were involved in the creation of the IGBP-DIS dataset.
4) Well organized, good data structure.
5) Someone once said "people ask for criticism when they really want compliments"; the authors of this database thoroughly deserve compliments!! It's amazing! Of course, I'd like several other factors to be included in the database, but you'd have to be brain dead not to be impressed by the potential of this concept. You'd also have to be completely lacking in appreciation of beauty not to be fascinated by the global NDVI images. The images are intriguing enough in themselves, but they certainly are also very useful.
With the African pilot project database I was able to map out a modified version of the Koppen classification. I could see how it compared with the standard Koppen and see that some locations were unclassified by the modified system. I could also look at the NDVI images to see if areas which were supposedly drought-free really were at particular times of the year. I'll be using my programs and the global database to do similar analyses for other areas of the world.
I purchased a CD-ROM reader just to see this database and I'm very pleased with the result.
6) I did not receive the database until a few months ago and have only now been able to begin using the data. From this limited experience it is apparent that this database will be very useful in our efforts to model trace gas emissions from soils and vegetation. Although many of these datasets are already available to us ..., it is more convenient to have the data on this CD-ROM.
7) .....I don't think that the full usefulness of this data will be known for some time. In the near term, I know of several people who are looking for this type of data, even though it is not for ecological modeling efforts. For my current interests, the elevation data is the most interesting and useful.
8) Highly useful as an integrated package. We put together a "home-grown" version of this type of global coverage, and it is useful to compare documentation and interpretations.
9) I am part of a larger ... Global Change program. The database is not helping me personally, as I am working on global aspects of carbon cycling and carbon cycling modeling. However, an integrated geographic database will be useful to our overall program which is working on modeling sectors of the U.S. including Alaska, Hawaii, and Puerto Rico.
10) This CD-ROM is just the type of database we need for teaching in our undergraduate and MSc level courses, and for some of the GIS research projects we would like to be involved with on a regional/global scale. The completeness and common basis of the raster data simplifies our tasks immensely.
11) This is an outstanding compilation of essential information for climate studies. I plan to make extensive use of it.
12) The database has been very valuable to me, in allowing me access to the Holdridge Life Zone approach description of the climate-vegetation relationship. The Holdridge approach seems to be the most readily available means of projecting vegetation patterns under global warming. Examining the included Holdridge layer helped me understand the advantage and drawbacks of using the Holdridge Life Zone approach for our study of biodiversity under global warming in the Pacific Northwest of the United States. A widely distributed, integrated database makes a lot of sense. It should allow workers from diverse discipline, institutions, and continents to compare results. I do have some misgivings about the global nature of the database though. The inevitable coarse resolution (at least of input data) will limit the usefulness for some regional and smaller scale studies pertaining to global change. I think consideration should be given to parallel efforts to produce national or bioregional databases modelled on GED, but at higher resolution.
13) As addressed in detail later, the present usefulness is minimal because of unacceptable access problems. The database is potentially very useful for scenario development and GCM evaluation research of [our program]. It also has potential applications for [our environmental monitoring, forest, and assessment programs].
14) The GED database is a useful product. All data are stored on the same medium (CD-ROM) with standard file and record formats. As the quantity of geophysical data continues to grow, it is important to store the data in standard formats and on easily accessed media such as optical disks. Since we did not use GIS software to access or process the data, I cannot comment on linkages between the GED database and GIS. We did not, however, find any of the IDRISI software that was provided useful.
15) The publishing of an integrated geographic database is timely with the development of systems to research and model the global impact to our environment that are occurring in this day and age. Locally, we plan to develop our own historic climatic database, and the effort by [our agency] to develop [a] database for management and modeling purposes should coincide well with the development of the Global Ecosystems database.
16) The database is quite useful, especially with the IDRISI interface. I have no complaints with the database itself. The organization of the documentation is good and the directory system and file naming conventions allow easy access to the data. I have found the Meteorological datasets quite useful in my research.
17) N/A
18) This is an enormously valuable effort. Already this database has substantially helped my research. I foresee that this database will be almost universally used and cited by global change researchers.
19) The database is exceedingly useful. I've been trying to obtain the NDVI data for years without success. These are all databases which are critical for global-scale research. Thus, having a documented common source of data is absolutely necessary for exchange of ideas. Also, it is wonderful for teaching classes and may attract MS.-level students into global research for their Ph.D studies.
20) The database is helping us in our research activities.
21) I think the database efforts are very useful. A major difficulty in developing spatial models is obtaining and controlling the quality of the necessary input data to run the models, especially if one must obtain these datasets from a variety of sources. Our simple ecosystem model... can use many of the datasets provided on the CD. With the CD-ROM data, we can compare the effects of using different input databases (Leemans and Cramer vs Legates and Willmott) on model results. The public access of this integrated geographic database should encourage additional modeling efforts by minimizing the time, money, and effort needed to obtain these initial datasets.
Another benefit of the database effort is the possible development of a "standardized" input dataset. It is often difficult to compare spatial models because the models have used different input datasets. By providing a "standardized" input dataset that is easily accessible to various modelers, models can then be compared based on differences of model interpretations of reality.
22) To the kind of work I am personally involved with, this database will be almost essential, especially if updates can be provided later.
23) It would be nice if every field were on the same grid. It would also be nice if the data were contemporaneous, i.e., January 1988, etc., for all fields.
24) 1. Extremely useful collection of information.
2. Each subsequent release should strive to greatly increase the amount of information included.
25) N/A
26) Very useful.
27) Yes, the efforts are useful, 3 years ago I was looking for data of this format to do a project with a quick turn around time and it was impossible, your database should reduce the waste of time associated with locating & working with large-scale environmental data.
[From cover letter:] The NOAA/NGDC Global Ecosystems Database is a useful and welcome package. The datasets included are certainly of value to research in global ecosystems. Some of the datasets are ones we planned to obtain from other sources (e.g., Zobler soils, World Data Bank II, Elevation and Bathymetry). Documentation of the datasets is good, both in content and format. In fact, we could make real use of the documentation even if we had to get the data from other sources. The inclusion of published documentation is quite useful.
Distribution on CD-ROM is a nice touch and useful. Organization of the CD is good. We operate in a UNIX workstation environment (DECs and SUNs), and we had no real trouble accessing the data files. I would have much preferred the actual data files be in the "lowest common denominator" of ASCII format rather than binary. However, with some work we should be able to access the binary data through GRASS (the GRASS appendix is helpful, but could give a bit more detail) or some other UNIX tool. The documentation of the binary format is of help here. Still, I would have been happier, and the data more useful faster, if they had been presented in ASCII.
The database will be useful in our research on global terrestrial vegetation. We would like to use it, and we look forward to the first real release.
28) The publication of this type of data is essential for many applications within the EPA. It provides a standard set of information on which further research can ba based. My major complaint concerns the accessibility to the GIS community. Although the PC based display programs worked... most scientific research using GIS is not done on PC platforms. When trying to determine how to convert the data to other formats for use in UNIX based GIS systems the documentation was inadequate. The database will only be useful if the data [are] provided in a useful format.
29) The Global Ecosystems Database is undoubtedly useful in our research activities. Its further development and increase in the number of datasets (parameters) will make it possible to widely use it in the field of hydrometeorology and ecology.
30) I believe that the publishing on CD-ROM in this way is very useful as a means of providing a baseline set of surface data. It helps to make the comparison of Earth observation data products between instruments and missions a realistic possibility.
31) Overall a very useful dataset; in particular inclusion of multi-temporal satellite data provides excellent opportunity to study global scale interaction between climate and vegetation dynamics.
32) N/A
33) N/A
34) 1. Selection of contents - The selection of databases - topography, temperature, precipitation, vegetation index, vegetation types, soil type -- seem to be good. [comment moved to 5]
2. Specification of grid type - Gridded datasets can be classified in two types .... One type may be called "grid point type" and the other, "grid box type". The former conceptually consists of point estimates at the cross points of x- and y- grid lines. The latter consists of some statistics, such as mode, median or mean, within boxes surrounded by grid lines. In addition, both types can be placed in two ways according to the position of grid lines. Usually grid lines are placed at integer multiples of the grid interval (or even number multiples of half the grid interval). Sometimes, however, grid lines can be placed at odd number multiples of half the grid interval, in order to match central point of "grid point type" data with that of "grid box data"....In general, the actual resolution of information is coarser than half grid interval, and the difference of grid types maybe irrelevant for processing of individual datasets. When making comparative studies, however, matching of grids is more serious. When making a database like this, there can be a choice. One can interpolate grid point type data to "odd" positions so that they coincide to center position of grid boxes. [Editor's note: The reader is referred to the explanation of registration and window coordinates in the User's Guide, for the conventions and definitions adopted in Version 1 of the database].
3. 10-min. datasets - I think it is good to include the datasets "expanded" to have the same spatial resolution as others. However, following points should be taken into account.
The procedure for making 10-min. grids from coarser grids should be documented. In the prototype manual, it is only documented for (3) only. For other datasets, it is just noted that they are derived by IDRISI program. At least, whether the value is just copied from the nearest grid point or interpolated from several grid points should be explained.
The original version should be included. For Legates and Willmott climate data (3), all the information can be incorporated in 0.5 degree version with smaller data amount than 10 min. version. [Editor's note: this would require either interpolation to "even" grid-box registration, or original data on "odd" convention and thus incompatible with the rest of the database. Expanded grids are no longer provided for software convenience, but only as needed to allow data intercomparability within the "nested-grid" convention.]
[Editor's note: comments about packed format omitted - packing is no longer used on the disc]
35) N/A
36) N/A
37) N/A
IA. WHAT SOFTWARE SYSTEMS DO YOU FIND MOST USEFUL WITH THE
DATABASE?
1) N/A
2) GRASS running on a Sun workstation. Could you link it to GMT either in xyz format or grd format (netCDF) files?
3) I think IDRISI is a good choice to present and use the datasets. PCs are cheap and quite common right now and IDRISI allows for immediate use of the data by colleges, universities and research institutes all over the world, including third world countries. However, there is a disadvantage to limit the data availability only to this specific computer platform. The research community especially wants to link global data to models and other large applications. PCs are less suited for such tasks due to memory and speed limitations. Although IDRISI is capable to export data to other formats, one still needs a PC to do such conversion -- [and a] program to read the data immediately and correctly from the CD-ROM (the compatibility with UNIX was no problem as stated in the introduction on page 3). To improve data exchange there surely would be a need of some conversion programs to read IDRISI data files and transform them to ASCII or binaries. The source-codes of these programs (Fortran, C, Pascal) should be provided so that they could be compiled and used on any computer system. This will immediately allow for the use of the global data on other computer platforms and therefore enhance compatibility and give the CD-ROM a larger audience.
4) A good layout. Looks to be very compatible or accessible to whatever the end user has or wishes to use.
5) I mainly use IDRISI and my own programs. I've been very impressed with IDRISI, which I hadn't seen before the African pilot project. I like the modularity of the system and the simple structure of the data files.
6) IDRISI
7) Clearly, the IDRISI GIS is most useful for an initial look at the data. In my case, I will be using the data with the programs I write.
8) GRASS 4.0, NCSA Datascope, XImage
9) We used the IDRISI program that was furnished to access the database.
10) We use a variety of proprietary raster handling GIS, mostly SPANS, but also IDRISI (very old version), and some vector types such as LaserScan's Horizon and Arc-Info. In addition we write our own on an experimental basis - usually raster - to provide functionality not readily obtainable from the proprietary ones. We have used GED with success with IDRISI and SPANS, and have imported it into some of our own half-baked (in the UNIX sense!) products.
11) We are developing our own to do hydrologic modeling. We also plan to make some use of a general-purpose GIS but haven't chosen one yet.
12) IDRISI. I had some difficulty attempting to transfer information from IDRISI to ARC/INFO for plotting.
13) Since the on-site reviewers were unable to successfully port any of the data in a form accessible for any of our software packages, I cannot answer this question. If access problems can be overcome, we plan to use the database in an ARC/INFO environment.
14) N/A
15) We were not able to access or view the data or structure due to lack of availability at this time of a CD-ROM drive. We plan to purchase one soon.
16) The IDRISI system was useful in the preparation of the data for further analysis. Unfortunately, it was a bit limited when it came to obtaining hardcopies of the displays.
17) In this review of the database I used IDRISI to examine the datasets. I also attempted without much success, to use the PC package MIPS to analyze the data. In the future I will be using IDRISI and GRASS in my analysis of the data.
18) I've imported all external databases into the BIOGIS system of Leemans and Cramer, which was tailor made for our work.
19) N/A
20) MS-DOS/IDRISI
21) Most of our modeling efforts and analyses have been PC-based. Therefore, IDRISI has been useful for displaying the datasets. To use the data with our models, however, will require transforming the data to a different data structure in which attributes are explicitly identified with a particular longitude and latitude. Once the data is transformed into this structure, we have used dBASE-IV to manage the spatial data; we have used SPSSPC+ and PC SAS to analyze statistics of the data; and have used GRAPHER and SURFER to display the spatial data.
22) Not really applicable as I understand. This disk provides data, not software. If the question concerns the usefulness of IDRISI in this context, then my comment is positive, because I am familiar with IDRISI. I am not sure, however, how big the problems with IDRISI are for users not familiar with the system.
23) IDRISI, and we can only display files, not access them. It would be better to be able to easily transfer ascii files to our computer system of Unix and Fortran.
24) 1. A text editor
2. IDRISI
25) ARC/INFO, because that's how we access geographic data at our site.
26) In-house software.
27) N/A
28) I don't believe the dataset should provide elaborate or even moderate GIS tools. Most researchers have tools for viewing, manipulating and analyzing spatial data and are more comfortable using their own tools. What they need is easy access to quality data and a convenient method of moving this data into their current analytical environment. Therefore I believe an extensive set of data conversion software would be an essential addition to this database.
29) IDRISI is acceptable to us.
30) We now use Suns almost exclusively: CD-ROM access is straightforward.
31) General image processing tools - we have been using IPW: Image Processing Workbench.
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) At present IDRISI, but we will add a UNIX system in the Fall.
IB. HOW IMPORTANT IS LINKAGE BETWEEN THE DATABASE AND GIS (AND/OR
OTHER SOFTWARE)?
1) N/A
2) Very important - allows early data manipulation.
3) For global change research this link is of uttermost importance. It permits displaying, linking and overlaying the data in many ways. Further, for users from other categories, it will give display and overlay capabilities, which makes this CD-ROM already more useful then for example the first part with environmental data of Times atlas. Without a GIS the data would be just a dull collection of [incomprehensible] digits.
4) Important to the novice but you can't get too coupled if you wish to stay flexible to other approaches.
5) I think this is very important, particularly as your target audience includes schools and colleges as well as researchers. Many researchers would be pleased just to obtain these data in digital form, but few schools would use the data if the link between the database and the GIS wasn't good.
I'd read that CD-ROM software was often unsatisfactory, so I've been pleasantly surprised by how smoothly IDRISI works with the database. That isn't to say a few things couldn't be improved. For example, several of the IDRISI commands were obviously designed to work well with floppy or hard disk systems and not the vast storage capacity of CD-ROMS. It would be good to be able to abort commands like LIST with cntl-C or perform selective searches, rather than to have to scroll through the entire contents of all active directories. Similarly, it might be useful to be able to abort long color jobs if they don't look interesting.
6) It is important.
7) For many researchers, this linkage can be quite significant as the raw data will be of little use unless they can display and manipulate it without having to spend a lot of time and effort writing software. However, it has been my experience that no one, or even a group, of programs can be expected to meet everyone's needs. Therefore, the availability of the raw data and reasonable documentation as to its content, format, and access is equally important.
8) Very important for our global nutrient cycling model studies.
9) We will be basing many ...information outputs on GIS. It is therefore important to us to link the database with GIS.
10) The IDRISI link has proved the easiest method to extract and display the GED. Input to SPANS enables quite complex modeling to be done. For teaching purposes the provision of a simple front end direct to IDRISI was very welcome. The only real requirement is to be able to import the data required onto disk in the correct form for the appropriate GIS. Working direct from the CD-ROM is unlikely to be a profitable route for repeat enquiries owing to the typical lack of access speed found on CD-ROM readers.
11) Not too critical, but nice to have.
12) Absolutely essential.
13) This linkage is very important for our intended applications. The existing linkage was extremely frustrating and extensive work, even using the GRASS interface, never provided successful access to the data. The data could be loaded into GRASS, but not be successfully displayed.
14) N/A
15) The linkage to GIS will have potential for our work in the future to spatially represent the data we collect/model, and also to integrate with other systems that are developing in that format. Currently, [our] GIS procurement is still in the pipe and not available to implement at this time.
16) I found the linkage between software and database essential, especially for intercomparing datasets.
17) I think the linkage between the database and other GIS packages is very important. Global (or continental) georeferenced data is very hard to come by (or too expensive).
18) For us not at all, i.e., the important thing is that we can easily read the files.
19) Not terribly; most people have a preferred database into which the software is translated (e.g., dBASE IV; Quatro-Pro, etc.). As long as the information in the separate datasets is transportable, the use is open.
20) The linkage is important for our research and initial stage of environmental modeling.
21) The linkage between the database and other software is very important. Without such linkages, the dataset becomes a set of pretty pictures to look at and perhaps gain some qualitative insights. With the linkages, the information in the database becomes available for many different types of analyses: data can be related more quantitatively to information in other spatial datasets or even in other non-spatial datasets; or the information can be used as input into models which may make the database even more valuable.
22) This is partly what I have answered above. Still, standardization of GIS has not reached the level where one standard would be satisfactory everybody. I think IDRISI provides reasonable links to the systems we use here. Essential is, that the system allows for the retrieval of raw data without deep knowledge of any one GIS system.
23) It provides a convenient way of displaying and manipulating the data. However, again we need to be able to take datasets easily to our own fortran programs.
24) Extremely important. In fact, more of the IDRISI Modules should be added, and a text editor should be added.
25) Its most important that it be stored in a common format. Most GIS's read the most common formats.
26) Not important.
27) N/A
28) Linkage is ESSENTIAL. Refer to previous answer.
29) Linkage between the database and GIS is essential, which should not however preclude from using other software.
30) Linkage is vital, especially to GIS on which the data have been developed. This enables those who wish to, to view data as seen by the originators.
31) This is important but not critical; more important is a simple, accessible data format so that data can be imported into local systems independent of specific hardware/software.
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Very - makes initial screening simple
IC. HOW IMPORTANT IS IT TO ENSURE COMPATIBILITY WITH A "LOWEST-COMMON-DENOMINATOR"
GIS PACKAGE (FOR EXAMPLE, IDRISI)?
1) IDRISI provided with the package was not a complete set. I should like [to] have [a] "shared" copy if it is not too late to request. (I received this prototype copy only in Feb. 1992)
2) Very important but requires compatibility with a non-PC environment.
3) Considering the costs in obtaining, training personnel etc. of most commercial GIS-packages, limiting the global CD-ROM to those packages, would be a severe limitation. Only large research and governmental institutes could afford these packages and use the global datasets. A broad distribution of the data is one of the aims of the release of this global data and therefore I think that IDRISI for the PC (and, if possible, GRASS for the UNIX environment) is the best choice in presenting compatibility of the datasets. Besides IDRISI is capable to import/export data from, respectively to other GIS-formats.
4) I think it would be very important.
5) Very important! By using a system like IDRISI, which is available on PC's, you make the database available to probably two orders of magnitude more users than if you used just GRASS/UNIX or similar. I think you'll be amazed by the analyses done by school students with this database.
Even more importantly by adopting a PC minimum configuration you make the information potentially available to students and users in the developing world, virtually none of whom have access to UNIX workstations. I say "potentially" because they also probably don't have access to CD-ROMs. It might be worth preparing some policy about circulating subsets of the information on floppies to people in developing countries. Could I extract data the final database and send it to a colleague in China or Africa? Would you want to purchase a copy of the CD even if they don't have a reader? This would be okay if the cost was reasonable.
You might also consider preparing a one disk "sampler" to load onto bulletin boards; it would be a good advert for IDRISI as well as the database. I'd also recommend sending a copy of the CD-ROM to [various magazines], when you release the final version.
6) It is very useful for users with limited GIS experience.
7) As with 1b above, that will depend heavily on the nature of the research and the fit between the GIS and the research needs. Although compatibility with the GIS is highly desirable, I would not eliminate data just on the basis that it was not compatible with a GIS. In the long run it is the data that is important, not the GIS.
8) Important, but also for Apple Macintosh GIS software.
9) I have as yet had no other experience with GIS. When we got a feel for IDRISI it worked O.K. However, the database will need to have Readme and Help files to make it readily usable with any compatible program package that it is set up for.
10) Very important indeed! Wide dissemination of the database will be predicated on easy and cheap use. Cheap but effective GIS on low budget platforms must be a large target market.
11) I think it is a good idea.
12) This depends on how widely distributed various types of GIS packages are. If "everybody" can get their hands on IDRISI, this makes sense. If, on the other hand, all serious consumers of the database have access to ARC/INFO (for example), distributing information on the "lowest common denominator" standard probably just slows down analysis.
13) This linkage is very important for our intended applications. The existing linkage was extremely frustrating and extensive work, even using the GRASS interface, never provided successful access to the data. The data could be loaded into GRASS, but could not be successfully displayed.
14) N/A
15) We feel the lowest common denominator is the best choice to ensure compatibility across systems that are in use.
16) The compatibility issue is directly related to the usefulness of the database. IDRISI should be either packed with the database or easily accessible to the user.
17) The database either needs to be in a low common denominator (like IDRISI) or it needs to have accurate conversion routines built into it to allow the export of the data to GIS and other packages. The current incarnation is perfect for this except for a few minor missing elements (discussed below).
18) I think this is important for many users, and I believe IDRISI is the right choice.
19) This is more important; most folks who could use the datasets in teaching, for example, are unlikely to have expensive GIS's, as are those in less-developed countries - it seems a small price to pay to maintain the IDRISI connection.
20) It's very important.
21) As a PC user and an IDRISI user, I think compatibility is very important.
22) see above
23) Zero.
24) Very important. Since that will almost guarantee compatibility with more sophisticated packages.
25) Not very.
26) Not important.
27) N/A
28) I don't support the concept of "lowest common denominator" GIS. There are now, and will be for quite some time, numerous GIS solutions available to analyze spatial data. What is needed for compatibility is a common data transfer data format for spacial data such as the Spatial Data Standards and Testing (NIST). Since this standard is still in review other commonly used spatial data transfer formats including DLG, TIFF, DXF, BIL, BIP... would be more useful to the GIS community than the IDRISI format supplied with the database.
29) N/A
30) Not sure - most GIS's seem to have foreign import options for "raw" data inputs, therefore, if the database format and structure are both simple then compatibility need not be a problem.
31) Not important; too much diversity in systems currently being used; rather, simple, accessible, well documented data storage format is critical.
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Very - for initial screening
- showing data to other users, quickly
But needs file format transfer programmes -- some are in IDRISI
already, others should be added.
II. ARE THE DATA ACCESSIBLE IN TERMS OF DATA STRUCTURES, FORMATS,
MEDIA, PROJECTION, ETC.?
1) N/A
2) Yes - but again would like easy access on the SUN/UNIX environment.
3) I had no problem in assessing the data files, both under the DOS and UNIX environment. Most data is in a gridded format and this can be transformed to any projection. I think that the data is very accessible on this CD-ROM.
4) Yes - good job.
5) Yes
6) Yes
7) The structures and formats used for the data on this CD-ROM are about as simple and easy to use as any I have seen. Where practical, simplicity is clearly a virtue. However, good data should not be excluded simply because it has a complex structure. Complex data just needs better, or more extensive documentation. I was most pleased to see that your data was unprojected (used standard Lat/Lon units). Most people with whom I come in contact want unprojected datasets. The quarter-inch backup tapes are currently in wider use and would also make good second choice. Given the availability of these low cost options, the traditional magnetic tape reels should be ignored in the future.
8) Yes, and I have included a procedure to access the data with GRASS 4.0. Feel free to share this with others, if it proves useful.
9) We found vector and window formats to be useful. We could have made useful printouts and used screen projection. I thought some of the data, especially on large-scale maps was questionable.
10) It is very instructive to compare access to this GED CD-ROM with the prototype issued for the world vector base (to be) provided by DMA called DCW.
The latter uses a highly structured file system intended for direct input to ARC/INFO, and is difficult and obscure to handle and slow to run owing to the repeated accesses to multiple directories and files on the CD-ROM.
The GED is simplistic but easy to use, fast of access, and simple to read and convert into alien formats. This is partly because it is a simple raster, but it could have been organized in some funny quad structured fashion to make life really difficult! Simplicity is always worthwhile when large volume data is being handled.
We did not test any projection handling, plate carreé is fine form for most of our needs.
11) Yes - very important to have data in lat/lon coordinates.
12) I found extraction of the data and use by IDRISI fairly straightforward, I did have a problem maintaining geographical reference points after extracting a regional subset of a raster file. Also, the lack of geographic reference the subset coordinators complicated transfer to ARC/INFO. Also, a means of subsetting vector files (e.g., from micro World Data Bank II) would speed processing enormously.
13) No. As mentioned above, rather than a "lowest-common-denominator" approach, compatibility should be maintained through the use of standard data structures and formats. For example, there are many industry standard image file formats, such as TIFF, BIL, BIP, and run-length compressed. However, should the database remain compatible with IDRISI only, then it is strongly recommended that software be provided to the reviewers to enable conversion of the datasets to GRASS format as well as to one/many of the industry standard image formats. Reviewers would not be expected to obtain this software on their own. It should be provided with the review package.
Another issue is time of access. This is really a hardware issue. To make the database accessible to the largest number of people, a PC platform is desirable. Unfortunately, this provides very slow access and transfer rates. for [our] general ... applications, the PC platform is probably unacceptable.
14) N/A
15) see 1(a)
16) One frustrating aspect in regards to the usability of these data is the lack of a suitable driver for the HP Laserjet Printer. Hardcopy of the results of analysis using IDRISI are necessary. Perhaps this problem will be solved shortly. The data format allows for easy access to the data.
17) There are a number of things to be done to improve the accessibility of the data:
A) The origin of the projection of the data needs to be given. I imported data from GOLSON.IMG into MIPS and converted it to an Alber's equal area projected image. While the image looked fine on the screen an overlay of the Great Plains states ended up in the Atlantic Ocean. It was difficult to establish proper georeferencing based solely upon the window information for the latitude and longitude.
B) A conversion routine needs to be added that would take the images in the database and convert them to one of the standard projections (i.e., Universal Transverse Mercator or Alber's Equal Area). While there are software packages available to do this, they require the data to first be in a vector format. I know that other raster GIS packages contain routines for projecting raster data. Being able to project the data would improve the display of it on the screen.
C) Much of my work deals with the abundance of a given species of bird at a given area. While I have a proper georeference point for my data, I could find no convenient way of identifying say, the GVI, for that data point. A VERY important (vital?) routine to add is the ability to import a list of points (in the proper projection) and then to export those points with the proper class identifier from the data. This would also greatly facilitate statistical analysis of the data.
D) CD-ROM, even though it's slow, is probably the perfect media for this project. As CD-ROM reader prices continue to fall, the data will become more available to more people. Other methods of mass storage simply aren't standard enough to be useful.
18) Yes, as far as we have discovered.
19) Just fine - no comments of value here.
20) Yes, the data accessible.
21) Data structure - As mentioned above, to use the data with our model will require a transformation of data structures. However, this change can easily be accomplished with a quick utility program.
Format - Personally, I like the format used to distribute the data because it emphasizes displaying the data. This format allows investigators to quickly evaluate the usefulness of the datasets for their research needs. (Especially when used with some of the palettes provided by the CD). In addition, this format allows modelers to quickly determine sources of problems encountered during model development.
Media - Although we did not have a CD-ROM player, I was able to borrow one to analyze the data and to transfer the data to a media (read/write optical) that we happen to use more regularly. Overall, I believe the CD-ROM is an excellent media for data distribution.
Projection - The lon/lat projection matched the projection we use for our modeling activities so the data were very accessible.
22) I expect CD-ROMS to be widely spread soon, therefore it is the most useful low-cost distribution medium. For us, any other medium would have been more costly in either investment of money or time.
23) Yes, however, user needs to be familiar with IDRISI to be able to access the files and display the images. Again, we need plain vanilla ascii data files.
24) Somewhat. I reviewed "(6) Matthews Vegetation, Land Use, and seasonal Albedo". I would have liked to output in ASCII file of the original data in the form:
LAT LON VALUE
. . .
25) Yes, we would prefer the data to be formatted for ARC/INFO, but then those without ARC would be out of luck, so this is a good compromise.
26) Yes - very good.
27) N/A
28) The data structure and formats used are not common spatial data formats (see 1c) and therefore the data is not easily accessible to most users in the GIS community. The CD-ROM media is a growing area and should cause no problem for researchers. Most GIS packages have projection algorithms to convert data from LAT/LONG to a variety of projections. Therefore, if the data is distributed in LAT/LONG, it should be no problem for the researcher to project the data to a desired projection.
29) Yes, they are.
30) Those examined were fine.
31) Mostly, however, in certain cases packing and coding format are not clear (see below); for the most part accessibility is good!
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Yes
III. ARE THE DATA ADEQUATE GIVEN YOUR RESEARCH NEEDS?
1) N/A
2) My needs are global modeling of the fluxes of trace gases - in this case these data are very useful - especially linked together on one CD.
3) Especially the climate and topography datasets are very useful for my research needs. These obtain continuous values and can be interpreted immediately. The vegetation and soil datasets are of lesser use, because these obtain classified data, which is more difficult to interpret. For example: what is the difference between potential vegetation, life zones, actual vegetation and land use. For inexperienced users it will be difficult to evaluate all these classes and therefore use the datasets in a competent way. Probably some guidance should be given in the documentation.
4) N/A
5) The data are useful, but researchers always want more and better data.
6) Yes
7) The coarse grain (5 and 10 minute) elevation data is good for a lot of general area work. However, it lacks the detail necessary for more precise work in limited areas. The one minute, three minute and 30 second topography data described on page 12 of the National Geophysical Data Center Solid Earth Geophysics catalog (January 1987) is highly sought after by the people with whom I am in contact. An additional CD-ROM with this data would be of considerable value and highly appreciated by a wide audience.
8) Not entirely.
9) We need to look at the data more closely to see if they can be used as a reference or base to build on.
10) Our research needs tend to be regional and continental rather than global. The GED is adequate but we are greedy.
11) I now have access to important dataset I expected would be hard to find. For now this is fine.
12) Inasmuch as our study concerns plants and animals with narrow distribution within the Pacific Northwest, in general the spatial resolution of the data is too coarse for our purposes. The Holdridge life zone data are probably adequate, given that the current state of climate change predictions does not allow any finer resolution of potential future climates.
13) I am unable to respond in detail to this question because of an inability to access the data. On paper, datasets such as the Global Vegetation Index, Monthly Temperature and Monthly Precipitation would be valuable as historical data for use by [our team]. Also, the thematic datasets would be valuable to [us] as reference data.
14) Not entirely. For our map of the Mid-West (see Fig. 1, attached), the level of detail of the geo-political boundary (vector) data were found to be adequate. For the Northeast (Fig. 2, attached), however, the level of detail was found to be inadequate. Important geographic features such as Martha's Vineyard and Nantucket islands are not well resolved. Other features, such as islands off the coast of Maine and the Connecticut River are missing entirely. Thus, the boundary data cannot be used to produce detailed maps for sub-continental sized regions.
15) There is only one dataset that we could find in the global climate database that would be useful, that is, the daily data compiled by Mueller (1982) based on the UK Meteorological Office series (1966-1983) and the World Survey Climatology (Landsberg, 1969-1981) with some other sources inspected as well. We could not find the locations of the 1057 stations that might tell us if we could use this particular data, but this is the type and frequency of data that we would use in our modeling work, and to build our database, compare data, or participate in the global database effort.
16) It would be nice if the raw data, i.e., individual data points used in the construction of these datasets, were accessible. It is understandable that these data could not be widely distributed due to space requirements. However, they should be available on a special order basis.
I would like to see satellite Geostationary (IR and visual) and polar orbiting data available on Optical Disk.
17) My research deals with scales of approximately .5 degrees on a side. As such, most of the data in this database is very useful. I am concerned however about the mixtures of scales within a given dataset. I found that the documentation was not very clear when this had occurred. As an example, The Olson World Ecosystem is given as having a 10 minute scale. In reality most of the data (with the exception of Africa) has a resolution of only 30 minutes and then it was converted to the 10 minute scale. I might be better to include the original data and a second set of the higher resolution data where available. (In this particular case having 10 minute data for North America would be fantastic!). Regardless, it should be made very clear in the documentation that there are mixed scales in the image.
One of the benefits (and one of the detriments as well) of the data is the scope of its coverage. It would be very useful to myself and other researchers to create a disk containing primarily North American data. As much as I would like to work globally, the data I need are just not available. While I can subset the data down to North America I think that a great many useful datasets may have been overlooked (both North American and European) because they were not global in origin. I realize that the scope of this project was to create a global database and you have done a very good job there. Perhaps you could either a) create a North America (or western hemisphere) specific dataset, or b) include other useful datasets that may not be global in coverage (perhaps in a separate subdirectory). Those datasets that were not global in origin could still be overlaid with images from the other datasets as long as they had been properly georeferenced and the proper window had been specified.
A priority for improvement would be to ensure that all of the maps were comparatively georeferenced at all scales. Another priority would be to try and get those images which are mixed in scale (i.e. GOLSON.IMG) finished at the same scale of resolution. It would be very worthwhile to have finer scale data available. Finally, I think that there needs to be some improvement in the acquisition software (as specified in the paragraphs above).
18) They are very useful, yet there are many desirable improvements.
19) The data available are excellent.
20) Yes
21) Our model uses data at various resolutions (although all input data are gridded to 0.5 degrees lon/lat) and produces predictions at the .05 degree lon/lat resolution.
22) Global modeling never will have the "adequate data". I believe, however, that several datasets on this disk are extremely useful for this purpose and, in fact, hardly available anywhere else.
23) No, operational analyses would also be useful, OLR would also be useful.
24) No
25) Yes
26) Some. (see enclosed notes)
27) N/A
28) N/A
29) They are, partially.
30) N/A
31) Yes, we have found this dataset to be quite interesting. In particular, we have found the database to be highly useful for explanatory analysis of relationships and interactions between vegetation and climate; we are also pursuing research related to map-guided detection of deforestation.
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Provides a first-order analysis to get started
IIIA. WHAT ARE THE LIMITATIONS OF THE DATA?
1) N/A
2) Available on PC only.
3) There is too little guidance in the documentation how to interpret and use the datasets. Climate, topography and coastlines are relatively obvious in what they display, but class data, like vegetation and soil characteristics, are more difficult to interpret. More guidance for users on how to use the different datasets would be welcome.
4) N/A
5) The coverage of some areas of the climatic interpolations could be improved. For example, Leemans and Cramer (1990) note that their interpolations have few data for Australia, south-east Asia and some other regions.
The International Soil Science Society is aware of limitations of the FAO/UNESCO soil data, especially for use with simulation models, so their SOTER (soils and digital terrain database) project is aimed at improving global and regional/continental coverage (more details from: V.W.P van Engelen, International Soil Reference and Information Centre, PO Box 353, 6700 AJ Wageningen, The Netherlands). However, the FAO/UNESCO data will remain the only global coverage for a long time.
I believe UNEP have digitized data for the more detailed (5000?) FAO/UNESCO map units (they kindly supplied me with these data for Africa), rather than the 133 soil units given in the database (e.g. Lc62-2bc would appear as just Lc in the database). The former data provide detailed description of soil conditions.
Box, E.O., Holben, B.N. and Kalb, V. (1989) Accuracy of the AVHRR Vegetation Index as a predictor of biomass, primary productivity and net CO2 flux, Vegetatio, 80, 71-89. Provides a very good introduction to the pros and cons of using the NDVI for ecological analysis.
6) N/A
7) As mentioned in the previous paragraph, the elevation data does not have sufficient detail for many applications. In my regular job I use Defense Mapping Agency CD-ROM based DTED on a PC. Your more detailed data nicely fills the gap between what is on this CD-ROM and that which is provided by DMA. For many applications the DMA data is too detailed as well as not being generally available to the public.
8) 1. For the purpose of coarse-scale global change studies, all
datasets should be provided at the 60 minute grid cell resolution.
2. Soil carbon and nitrogen datasets (Post et al., 1985,
Nature 317:613-6) would be useful additional layers.
3. The gmveg datasets (especially cultivation intensity) is
probably badly outdated. A new assessment based on global coverage (Landsat) imagery would be much preferred.
9) Probably, the most serious limitations are the broadness of the scales.
10) Our wish list would definitely like to see all data collected on 2 minute (as with the FAO), or 5 minute at worst. But we will happily take what we can get. We are worried about the problems of nesting because of the scale implications, but would rather have non-nesting stuff as to more detailed level than pure data at 1:10m.
11) I need a finer resolution database. Orographic precipitation modeling requires 1km - 2km (1min or less resolution).
12) Primarily spatial resolution (see above). The means of combining life zones for the aggregated Holdridge life zones dataset was not specified. This limited usefulness of that dataset.
13) N/A
14) N/A
15) N/A
16) The analyzed products are only as good as the raw data used to produce them. I was happy to see that the error analyses produced by Legates and Willmott was included.
17) N/A
18) Ground-based climate data as they stand are rather problematic - they have uneven data coverage and quality, and don't include e.g., radiation and extreme event. The NDVI period covered is fairly short (compared with e.g., the climate data). The vegetation datasets have numerous limitations e.g., doubt about many of the sources, doubt about the method used by Matthews to infer natural vegetation, and chaotic documentation by Olson.
19) None
20) N/A
21) We have experienced tremendous problems using Matthews vegetation dataset (1o lon/lat resolution) in conjunction with Legates and Willmott (gridded to 0.5o lon/lat). We believe the source of the problem is in the difference of resolutions of the two datasets. When we use vegetation datasets digitized to 0.5o lon/lat we have not had nearly as many problems.
Soil texture is another input into our model. We have been using texture from the FAO map gridded to 0.5o lon/lat. Your Global Database only has Zobler gridded to 1o lon/lat. This resolution difference might also be a problem for the reasons described in the previous paragraph.
22) This would be major topic of study. Some short comments:
- the climate data fields (2) and (3) are all nice, "smooth" data fields, which obviously reflect the true conditions very differently in different regions (due to shortages of input data). The user has no chances to estimate the spatial pattern in the reliability of the data. More important, the resampling procedure from the Legates and Willmott grid to the standard grids in this database, seem to be a dubious and "dangerous" procedure to me. Ideally, the gridding from original raw weather station data to whatever desired output grid should be performed in one single step (see pt.4 for further comments on this).
- none of these datasets can ever be expected to be "final", but some of the limitations in the Olson dataset (4) can be derived indirectly from the upgrade documentation in this volume.
23) Only a few meteorological parameters are on disk.
24) I would like at least three times the number of categories for the albedo data.
25) N/A
26) Resolution - often too coarse. Patchy data (large blocks w/same value)
27) N/A
28) N/A
29) The data application is limited by the scale of time averaging, the inadequate amount of geophysical parameters, included in the database.
30) N/A
31) Obviously, a larger time series of satellite data would be desirable. Also, precipitation and temperature data contemporaneous with satellite data would be very useful. Consistent spatial resolution would be useful (e.g., use UTM projection)
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) #2 Soils -- need a water-holding capacity and other quality interpolations - such as the GISS GCM routine
#1 Climate -- monthly time series and a consistent averaging period
required -- IIASA data is length of record
IIIB. WHAT ARE THE PRIORITIES FOR IMPROVEMENTS?
1) N/A
2) Access on other platforms in other formats.
3) The documentation could be improved considerably by adding summaries, scope and limitations of each dataset to the "dataset description" pages. I believe that this should have the highest priority.
4) N/A
5) Improvements in coverages of temperature and rainfall interpolations. Inclusion of full FAO/UNESCO soil unit data.
6) additional datasets
7) For my areas of interest, your more detailed topographical data is top priority. Since the data already exists, it would appear that this addition to the CD-ROM based database might be made with minimal expense and effort.
With respect to the current data, it appears that some of the datasets have voids in them. This impression is the result of a limited look at the data with the IDRISI GIS. If such voids exist, their elimination would appear to be a priority area for attention.
8) 1. Updated land use coverage.
2. Soil attribute data layers.
9) Research is needed to build reliable sector data banks over a period of years.
10) Firstly a more detailed spatial resolution, and secondly a wider range of topics.
11) Higher resolution terrain data.
12) These are perhaps software concerns, but I would suggest the following as priorities for increasing usefulness.
1. build in a means for maintaining geographic reference of datasets through various transformations including reclassification, subsetting, and raster-to-vector conversions;
2. provide a means for subsetting vector files
3. enhance the range of hardware (i.e., printers, plotters) compatibility for IDRISI's "plot" module so as to facilitate producing hard copy.
13) N/A
14) N/A
15) N/A
16) 1. Error analyses for the other products included.
2. Accessibility of raw data.
17) A priority for improvement would be to ensure that all of the maps were comparatively georeferenced at all scales. Another priority would be to try and get those images which are mixed in scale (i.e. GOLSON.IMG) finished at the same scale of resolution. It would be very worthwhile to have finer scale data available. Finally, I think that there needs to be some improvement in the acquisition software (as specified in the paragraphs above).
Olson World Ecosystems - (A few dataset notes - to me this is the most useful dataset on the disk, it would be helpful if Dr. Olson could be persuaded to complete the 10 minute updates for the rest of the globe, or at least the Western Hemisphere). Why are so many categories not used? I think it would be better to eliminate the not used categories to reduce the amount of searching necessary to find a category corresponding to an area of interest. If color choice is a problem then perhaps a special palette could be provided. What documentation? Except for a brief, useful, review of how the dataset was updated the documentation is nonexistent. What are the classifications based on? How have the classifications been used (references?)? The reference from Milleman and Boden, 1985 should be included as part of the supplemental literature (especially as it is not readily available in most libraries). [Editor's note: This is an incorrect reference that was mistakenly used in the Prototype release -- the correct reference is Olson, Watts, and Allison, 1985, ......]
18) - work with GCM/weather prediction modelers to provide long-term "data" on all key variables from model output generated shortly after updating.
- include a longer period of NDVI data, and also monthly and annual averages over the whole period.
- work with IGBP-DIS toward a satellite-based global vegetation dataset.
19) 1. other datasets as these scales
2. plus remote sensing at finer scales
3. inclusion of hardcopy of each mapped variable for comparison to electronic ones on screens to confirm data are correctly displayed.
20) N/A
21) For our research purposes, the availability of improved NDVI datasets is highest priority. Next, the availability of time series maps of land use and climatic data is highly desirable.
22) - improved gridding techniques for continuous variables, such as the climate variables
- improved world ecosystem database
- improved (corrected) topography and bathymetry database
- new soil texture database
23) Plain ascii files for transfer to a wide variety of systems.
24) 1. Ease of coordinate/data extraction
2. A note that the "Menu.exe" utility uses the default directory c:\IDRISI\
3. Acquisition software on this CD-ROM
25) Agricultural practices and other historic databases.
26) same as 3a.
"resolution - often too coarse
patchy data (large blocks w/same value)"
27) N/A
28) N/A
29) Database expansion on the basis of comments in 3a.
30) N/A
31) - Longer time series of GVI
- What about including data from Earth Radiation Budget Experiment (ERBE?)
- More current climate data
- Sea Surface Temperature?
- How about including point data from which climate data/images were interpolated?
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) (Probably outside your scope of work:)
- socio-economic data -- population, land use, agricultural zones, cities, etc.
- analytical routines -- aridity index (as in UNEP Atlas of Desertification), agroclimatic zones, vegetation biomes -- could all be set up in IDRISI (?).
- GCM scenarios of climate change
IV. DO YOU HAVE DATA CREATED OR OBTAINED THROUGH YOUR OWN RESEARCH
THAT YOU WOULD LIKE TO CONTRIBUTE FOR THE NEXT RELEASE? PLEASE
DESCRIBE.
EDITOR'S NOTE: Because of the sometimes sensitive nature of data
availability issues, answers to this question are provided only
in summary form, in the main body of the User's Guide. The project
will pursue and inventory relevant data sources, as availability
is confirmed.
V. PLEASE PROVIDE SUGGESTIONS FOR OTHER DATA THAT SHOULD BE
OBTAINED OR FOR CHANGES IN THE CURRENT Datasets. PLEASE BE AS
SPECIFIC AS POSSIBLE, GIVING SOURCES OF DATA IF KNOWN.
1) It is possible to add information such as;
Nature: Volcanoes, large faults, historic landslides
Human structure: Dams, mines, large cities, observatories, and their names; and population density
2) Complete model runs with all fields from global climate simulations would be very useful eg., NCAR, 1XCO2, 2XCO2 runs these could be updated regularly.
3) The US-EPA and NCAR have compiled a tape with several GCM climate change scenarios. It could be very useful if the main parameters, temperature and precipitation change, are included in the global datasets. This will make the CD-ROM more useful for global change research.... The reference could be (I am not completely sure): Jenne, R.L. 1989. Data from climate models: the CO2 warming. National Center for Atmospheric Research, Boulder, CO.
The "Trends 1990" - data is somewhat outside the scope of this CD-ROM, but could be useful for global change research. If there is space left the datasets could be included, but most of them are not geo-referenced and could not be formatted for IDRISI. The correct reference is: Boden, T.A., Kanciruk, P. & Farrell, M.P. 1990. Trends '90: A compendium of data on Global Change. ORNL/CDIAC-36, Oak Ridge National Laboratory, Oak Ridge, Tennessee.
The Zobler and UNESCO/FAO soil database are the only digital ones currently available, but their quality is limited. I have had correspondence with the FAO and learned that they are working on a new better digital implementation of the proper FAO-soil map of the world. The CD-ROM should include this database instead of the other two if possible. The FAO-database should be available by autumn....
I find the soil databases useful, but for modeling purposes I need databases with more relevant parameters then soil, texture and slope class. For some purposes I have created some derived datasets with different soil characteristics, like water holding capacity, fertility and soil depth. Because of the simple correlation schemes that I have used, I do not dare to present these derived datasets to others, but I would promote to find or create global datasets on these properties, created by experts in this field.
Only climatic averages are included in the dataset. It should be nice to include also some climatic trends and series based on series of monthly values. The best dataset around is the one compiled by Bradley et al. (1985). I know that recently there has been an update of this database.....
Reference:
Bradley, R.S., Kelly, P.M., Jones, P.D., Diaz, H.F. and Goodess, C., 1985. A climatic data bank for the northern hemisphere land areas. 1851-1980. DOE Tech. Report No. TR017, US Dept. of Energy, Carbon Dioxide Research Division, Washington D.C.
4) N/A
5) A GCM-developed climate change scenario for temperature and precipitation in about 50 years time would be interesting e.g. as used by,
Emanuel et al. (1985) Climatic change and the broadscale distribution of terrestrial ecosystem complexes, Climatic Change, 7, 29-43.
I guess one of the GCM groups would be able to supply a more reliable prediction than the one Emanuel et al. used years ago.
Interpolated data for monthly evaporation (Penman-Monteith estimate) for the whole world would be very useful. The estimates of evaporation based on just temperature and precipitation, as used by Holdridge, are not very reliable. Better estimates of the water balance would improve predictions of ecosystem distribution. Monthly interpolations for solar radiation could be done at the same time (see Hutchinson, M.F., Booth, T.H., McMahon, J.P. and Nix, H.A., 1984, Estimating monthly mean values of daily solar radiation and the more commonly available sunshine hours measurements). These factors along with monthly maximum and minimum temperature (and the existing precipitation relationships) provide the basic data necessary to run simulation models see for example,
Booth, T.H. 1991, A climatic/edaphic database and plant growth index system for Africa, Ecological Modeling (in press).
In the longer term it would be good to interpolate gamma parameters (see
Samani, Z.A. and Hargreaves, G.H., 1989, Applications of a climatic database for Africa, Computers and Electronics in Agriculture, 3, 317-325), so users could generate synthetic sequences of rainfall and link the data for specific locations to agricultural models, such as those being developed by International Benchmark Sites Network for Agrotechnology Transfer (Dept of Agronomy, Univ. of Hawaii, see their DSSAT model system). In the shorter term, people could use the actual time series data in the World Weather Disc CD-ROM (Weather Disc Associates, 4584 NE 89th, Seattle) to generate time series data for any location on the sample grids (i.e., by comparing means at locations with actual data with those sample locations and calculating a correction factor).
6) It would be very useful to have more detailed land use data and natural vegetation types. For example, regional vegetation in the U.S. in about 100 categories. (e.g., Oak-Hickory forest, bluestem prairie).
7) [comment moved to 3b]
8) same as 3b
1. Updated land use coverage.
2. Soil attribute data layers.
9) Thought should be given to inclusion of maps that track generation of greenhouse gases. The database could be improved through addition of parameters more directly related to sources and sinks of greenhouse gases and quantitative indices of radiative forcing.
Information on CO2 generation from depletion of forests in tropical countries and consumption of fossil fuels in industrial countries would be useful. Data on tropical forest depletion is available from FAO. (Food and Agricultural Organization of the United Nations in Rome).
10) Presumably you have had talks with DMA about DCW when it hits the public domain. This might provide some of the extra infill information for relief? How about including hazard and other data (possibly already in the World Data Center) - not necessarily on a raster basis eg: global earthquake/seismicity files.
Is there any general geological or geochemical data available on a raster basis? This could be useful? Air pressure? Air quality? What about economic minerals maps?
I realize all these above would require effort to get from paper map form into classified raster, but they would be valuable. The present GED appears to have as its major constituents input from classified satellite imagery, plus terrain, submarine, and soils information from other sources. Perhaps there is a case to be made for serious nested raster import of paper mapping products?
11) Drainage connectivity array consistent with the terrain data.
12) You might consider including some global change projection datasets, e.g., GCM output under doubled CO2, or Holdridge life zone maps based on the GCM output. The advantage of this would be a common basis for a variety of "what-if" exercises by a variety of workers. The drawback would be the implicit 'stamp of approval" on the global change scenarios included. Whether the benefits would justify the controversy that might ensue would require some careful thought and discussion to determine.
13) N/A
14) N/A
15) a) As mentioned in response (3) Daily means, max, min, etc. for the Southeast Region, specifically, the closest official NOAA weather station.
b) Monthly global vegetation index. Monthly resolution is not adequate for our modeling purposes. Our interest would be in identifying when the phenological season begins by latitude and longitude. Weekly would be satisfactory. Sources include Gutman, G. in Appendix A [of the GVI User's Guide] - Bibliography of GVI Research Reports.
c) The monthly datasets of temperature, precipitation, and cloudiness have some utility to our research program. Derived indices of moisture demand; PET, AET, etc., are useful. Weekly or daily values are necessary for serious modeling of vegetation functioning.
16) Satellite infrared and visual data should be available on a grid which is determined from space requirements. The data should have as high a spatial resolution as possible. The international Satellite Cloud Climatology Project [NASA] makes these data available...
17) Some of the data that I would like to see available in the next release include:
a) The various GCM models
b) Changes in the Holdridge Life Zones under the various GCM scenarios (possibly available from NCAR)
c) Monthly evapotranspirations.
d) Monthly moisture indices (Thornthwaite) along with a map of Thornthwaite's classification of moisture requirements).
e) For North America - Kuchler's potential vegetation, USGS digital line graph (DLG) vector data and possibly some digital elevation model (DEM) data.
f) Finer scale (1-4 km) AVHRR data.
g) When available, more time series data.
18) Why not include paleo datasets (e.g. for 6000 and 18,000 yr. B.P.) for model validation purposes. For example, DOE has a dataset of global lake - level changes. Other such globally standardized datasets may come into being soon. Suggest liaison with NOAA [Paleoclimatology Program].
See above my comments on previous page. For the concept of using weather prediction models to generate grid-based, physically consistent data contact. John Kutzbach at the Center for Climatic Research, Madison. This is one way to provide data important for ecological and hydrological modeling e.g., winter minimum temperature, net radiation, storm frequency.
19) Global Land-Use Patterns: through my IIASA project, David L. Skole of University of New Hampshire, Center for Earth, Oceans and Atmosphere, obtained and now has digitized, Soviet maps of global land use patterns. Because climate along with landforms constrains land use, these and similar data are very important sets for inclusion in your database.
Climate Change Scenarios - the current set of EPA-distributed climate scenarios from GISS, GFDL, CCC, etc. could be added profitably, with suitable cautions about their limitations. Especially useful would be intermediate data not usually available on daily temperature or precipitation extremes as model output.
Atmospheric chemistry data - I am unaware of any dataset which could provide global distributions of atmospheric NOx, SO2, etc. Perhaps enquiry among the users of your CD would bring out this useful database.
20) I would be delighted to include the collection of the data about paleo-environmental change for the last 130,000 BP. At least 125,000 18-20,000, 10,500, 8000, 6-5500, little ice age, little climatic optimum of holocene. The data about anthropogenic climatic change also are very important.
21) We would be interested in another global dataset of cloud cover (0.5o lon/lat resolution) that is based on satellite data.
In our quest to improve our vegetation data, we have been "hand digitizing" the following vegetation maps for our model. You may be interested in obtaining the following datasets:
Belov, A.V., I.I. Bruks, E.K. Fedozova, S.A. Gribova`, T.I. Isachenko, N.M. Calibernova, Z. Y. Caramisheva, G.S. Catenina, G.M. Ladygina, V.V. Lipatova, N.R. Lifvinova, N.I. Nicolskaya, E.I. Racheovskaya, I.N. Safronova, V.N. Haamtsov, T.C. Yurcovskaya, I.S. Ilina, Y.S. Prozorov. 1990. Vegetation of the USSR. Moscow State University, Geography Department, Moscow USSR.
Hou, H.Y., S.C. Wang. 1979. Vegetation Map of China. The Laboratory of Plant Ecology and Geobotany, Institute of Botany, Chinese Academy of Sciences, Map Publisher of the People's Republic of China. Beijing, China.
Joint Federal-State Land Use Planning Commission for Alaska. 1973. Major Ecosystems of Alaska. U.S. Geological Survey, Fairbanks, Alaska.
Kuchler, A.W. 1964. Potential Natural Vegetation of the Conterminous United States. American Geographical Society, New York, New York.
Rowe, J.S. 1972. (Reprinted 1977) Forest Regions of Canada. Publication No. 1300, Department of Fisheries and the Environment, Canadian Forest Service, Ottawa, Canada.
UNESCO. 1981. Vegetation Map of South America. UNESCO, Paris, France.
White, F. 1981. Vegetation of Africa. UNESCO. Paris, France.
22) I have heard of US activities to create an improvement for the terrain dataset (I expect you have better and more specific information about that).
A most important type of data I would like to see on this disk is anomaly fields derived from GCM's, similar to the standardized files available from NCAR. These files should be at both, GCM resolution and gridded to some of your grids.
23) Numerical weather prediction analyses data.
Operational NMC analyses that give 3-dimensional picture of atmosphere structure. Contact ... NCAR and ... NMC/CAC; Outgoing Longwave Radiation, ... NCDC.
24) The GCM output mentioned in the manual is a good idea.
A collection of the satellite imagery of earth from the Spot and Landsat would be extremely useful (Available from EOSAT ....).
25) N/A
26) Substitute full CIA World Data Bank II for micro version. I think this has greater resolution and more attributes. Contact NTIS ....
You might want some data from the DMA's Digital Chart of the World Project. Contact ... DMA....
27) N/A
28) N/A
29) ...Hydrometeorological and environmental datasets which have been created or are being created in different organizations in Russia and outside it. The sources can be specified...
30) Seasurface temperature
AVHRR
GHOSTA (UK Met. Office Database)
Atmosphere data - mean values
Total water vapor (SSM/I) - already done by NSSIDC
31) Continental-scale higher resolution NDVI data from NCAR or EROS data center.
Time series of seasurface temperature (NOAA)
Sea Ice? NOAA?
Chlorophyll concentration from CZCS - NASA
32) N/A
33) N/A
34) I think inclusion of variables related to surface energy and/or water balance, e.g. net radiation, evaporation, runoff etc. can be useful. However, since there are perhaps no datasets ready to be included, they would be put off to later editions. I wonder why Wilson and Henderson-Sellers' vegetation dataset is not included [Editor's note: we were awaiting copyright release]. It is different in philosophy from Matthews digitized potential natural vegetation, while Wilson did existing vegetation. Both approaches have their merits and demerits. Wilson's soil dataset is derived from FAO/UNESCO soil map, so it may not be necessary now when Zobler's dataset is included.
35) N/A
36) N/A
37) Our goal (see the attached project description) is to compile
the FAO agroecological zone data (improved to worldwide coverage)
and model on a CD-ROM. This would integrate the natural resources,
agricultural suitability and socioeconomic interpretation into
one database/GIS that users around the world could test -- either
globally or regionally by filling in finer-scale data and analyses.
VI. IS THE YEARLY RELEASE SCHEDULE REASONABLE, GIVEN THE PACE
OF GLOBAL CHANGE RESEARCH AND/OR YOUR RESEARCH?
1) N/A
2) The latest release should be available over the network with annual updates of the CD-ROM.
3) I do not think that a yearly release is practical. Updating of most datasets will take a longer periods, especially because the creation of datasets is not the main aim for global research, although often on the agenda. This can easily be determined from the documentation. Many papers describe the datasets only in summarizing terms and emphasize on the application. If each release of the database should be an improvement over the previous, a biannual release schedule should be more feasible.
4) We're not that familiar with your users but a yearly publication of a CD is in concert with some of the things we're doing.
5) Yes
6) Yes
7) For the data on the disk a yearly update would appear adequate. For the additional topography data mentioned above, the sooner the better.
8) Obviously, the sooner the better.
9) I would consider it reasonable.
10) Yes
11) Yes
12) Yes
13) Yes, except in the case of annual reports associated with [our] program. An annual report is required 9 months after the close of sampling (July). We already have some problems obtaining current climate data in time and I imagine the lag would be even greater with this system. For more traditional research and development activities associated with [our programs], the annual time frame should be adequate.
14) N/A
15) Yearly release of updates to the database are reasonable based on the need for data of high quality, refinement, and release on a ROM disk.
16) The yearly release schedule is adequate as long as some flexibility is included to account for variability in the production of new and important datasets.
17) A yearly release schedule is reasonable only if the data significantly changes. Rather than include the same data each year it would be more useful to only include corrections and new datasets in each yearly installment. Along the same lines it is very important to keep researchers appraised of errors found in the database. Perhaps a dial in bulletin board would be more cost efficient than periodic mailings.
18) Yes - we don't need to update too often; there is something to be said for the use of standard databases that are "valid" for a reasonable time, e.g., 1yr.
19) Seems very reasonable although if you are funded well enough, you could send out a quarterly/monthly newsletter which describes B-test datasets recently available, for those who cannot wait!
20) It's reasonable, at this stage.
21) I'd be concerned about the quality of the data if I was to develop datasets for annual release. However, with several groups working on global datasets, I think a yearly release schedule is reasonable if different groups contributed to a specific release. Indeed, at a later date, it may be desirable to have more frequent releases.
22) Yes
23) Yes, although it would be nice if it were more often. Also, these files really should be temporally consistent.
24) Yes
25) Yes, to ensure the high quality of the data and to let it be reviewed.
26) Yes
27) N/A
28) Yes
29) The yearly release schedule is reasonable.
30) I believe that yearly release is probably the maximum rate at which these data should be released. The processing of checking data formats etc. and validating actual values is necessarily time consuming.
31) Yes - it is more appropriate to provide data of dependable quality with good documentation, than to release data prematurely!
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Yes -- see [question] #12.
VII. IS THE DOCUMENTATION HELPFUL, COMPLETE, ETC.? (PLEASE
ATTACH COPIES OF THE DOCUMENTATION SHEETS FROM THE MANUAL WITH
YOUR EDITS, IF POSSIBLE).
1) N/A
2) Very satisfactory.
3) The documentation gives currently the most important information. I would like, however, the addition of a short summary and some notes on the applicability of the database. The summary (c. 100 words) would describe the data and would allow for a faster evaluation of the content then by looking at the proper documentation or requesting the additional literature references. The extra notes would summarize the scope of the data, its limitations and applicability.
Each contributor of global datasets should be urged to fill in a standardized information sheet with room for all important information. If the contributor has not filed such information sheet, the dataset should not be included.....
- The correct reference to the IIASA climate database is now:
Leemans, R. & Cramer, W.P., 1991. The IIASA database for mean monthly values of temperature, precipitation, and cloudiness on a global terrestrial. Research Report RR-91-18 November 1991, International Institute of Applied Systems Analyses, Laxenburg, 61 pp.
The reference Solomon and Leemans (1990) should be moved to additional references. It presents an application and does not need to be included as primary reference, but as additional reference. Another additional reference could be:
Prentice, I.C., Cramer, W., Harrison, S.P., Leemans, R., Monserud, R.A. & Solomon, A.M. 1992. A global biome model based on plant physiology and dominance, soil properties and climate. J. Biogeography (in press)
......I find the documentation of the Olson dataset insufficient. In my global change research, I found this the best vegetation database currently available, better than, for example, Matthews, because very clear descriptions of each vegetation class are given in the accompanying report (Olson, J., Watts, J.A. & Allison, L.J. 1985. Major World Ecosystem Complexes Ranked by Carbon in Live Vegetation: A Database, NDP-017, Oak Ridge National Laboratory, Oak Ridge, Tennessee.) This reference should be included in the documentation as primary reference and not as additional, and if possible, also a reprint of this report.
The soil databases (Zobler and UNESCO/FAO) are all based on the FAO soil map of the world. I find it essential to include the original reference of the map the documentation. This reference is:
FAO/UNESCO 1974. Soil Map of the World, 1:5,000,000. Food and Agricultural Organization, Paris, France.
4) Yes, quite thorough for a beta test.
5) Yes. I've only had a couple of minor problems so far.
When I first tried to use ACCESS I received an "Out of environment space" message. I had to trim my existing long path statement. I believe there's a DOS command to increase the environment space, but I've not had time to track it down.
SHELL C:command.com /E:400 /p increases environment space to 400 bytes from default of 1606.
I thought the description of the use of SUBSET could be improved by giving an example of its use. If I want to just look at Australia do I have to extract all the relevant data using SUBSET or is it possible to change the default window that COLOR uses, so that it would show Australia first rather than North America? It would be a good feature to add if it's not presently possible. Incidentally, the new windowing feature of COLOR is excellent.
6) N/A
7) I found the documentation to be generally complete and appropriate. Some of the copied material was difficult to read and some of the graphics lost much of their usefulness due to blurring of the copies. Where possible, you might ask the contributors to resubmit cleaner originals. The IDRISI and World Data Bank documentation was probably the least appropriate and complete. My full comments on IDIRISI and World Data Bank are provided below. To reduce your overall production cost you might consider putting as much of the documentation as possible on the CD-ROM. If maximizing the potential use of this excellent data is a goal, then keeping the total distribution costs to a minimum would appear to be a necessity.
I hope that the IDRISI documentation extract provided in Appendix A is not representative of the full documentation, especially if the intent is to use IDRISI as some form of standard GIS for working with the data at the high school level. Although the various modules (programs) provide considerable capabilities, it is unclear from the extract how the capabilities are integrated. Although the menu program provides an interface to each of the programs, it does nothing to integrate the capabilities. Based on what is provided, it is my impression that it will take a fairly computer literate and experienced teacher to figure out how to use the software/data combination. The teacher will then have to figure out how to package the programs so that the students can make use of the capabilities. The general syntax and contents of the documentation is typical of a computer technical reference manual - not something which will be effectively and efficiently used high school teachers and students....
In all fairness, I must admit that even this partial IDRISI GIS has been of considerable help in looking at the data and gaining an appreciation for the considerable value of your efforts.
If there will be additional reviewers who will be working on this project in the future it might be useful to do some rework on the extracted documentation and build some step-by-step tutorials to get them started and save some time.... The value of the data you are assembling should not be diminished by the IDRISI documentation.
8) In general, documentation is adequate.
9) The documentation is helpful and complete, However, there needs to be better explanation, with examples, to make the data from the CD-ROM more readily accessible and user friendly.
10) Most impressed with the documentation. The only major problem is that for non-IDRISI users there may well be some difficulty in getting going. The IDRISI section does not indicate how a sensible path through the various modules might be chosen.
11) Yes, but I'm just starting to use it.
12) As mentioned above under 3a), description of the method of combining life zones for the aggregated Holdridge life zone datasets is needed. Also, some discussion of the similarities and differences between the two monthly weather datasets would be helpful.
13) No! This, along with the "lowest-common-denominator" issue were the strongest comments from my technical reviewers. For example,
"Specific documentation on the structure of these image files does not exist, e.g. data types are defined as 'byte' and file types are defined as 'binary'. I did find one reference that some grid files (only the 8-bit raster files), are identical to GRASS cell files. I did attempt to manipulate a ten-minute GVI file into an appropriate GRASS cell subdirectory, but I was not successful. In Appendix B, the documentation reports that 'import of the data into GRASS is relatively easy...' however, the special software is not provided and more detailed instructions are necessary."
14) Some of the documentation is confusing. Descriptions of the datasets found on the DATA FILE STRUCTURE pages are terse. At first we did not realize that most of the data record formats described by documentation in sections (1) to (11) are relevant not to the GED files, but to earlier versions of the same data. Almost all of the information that one must know to access the GED versions is found in the GED Introduction. Since, when first reading the documentation, one might easily pass over these few pages of information, a note or statement on each Dataset DESCRIPTION or DATA FILE STRUCTURE page that directs the user back to the GED Introduction might be helpful. (See notes on attached copy of data file structure page.)
Some of the additional information provided in sections (1) to (11) discussed the sources and history of the datasets. This information is important and should remain part of the GED documentation.
15) See pages 8-10 for example. It is difficult to tell which datasets are of the same format/file if they are separately collected and formatted. Are these datasets in some kind of relational database or separate files? We may have been able to ascertain this a little better if we had been able to view the CD-ROM disk.
16) The documentation is essential. It was quite complete and useful. The photocopied journal articles were very light and hard to read in some places. They should be reproduced using a text editor.
17) General notes concerning the documentation. The documentation should continue to be released as a hardcopy document and should also be broken up into two volumes. The first volume should contain all of the introductory material, dataset descriptions, and an overview of the data (probably synopsized from the original sources). The second volume should contain the supplementary material. In particular, a copy of the actual paper that describes that dataset (not just the reference) would be very helpful. Additionally, one or more papers that demonstrate the use of the dataset should also be included. I realize that in some cases this represents a significant effort to secure copyright approval. However, if the manual was well done it could probably be published as a stand alone book as well as supplemental information for the database (I'd certainly read it).
18) Mostly yes, but I am dissatisfied with the Olson and Matthews documentations. Olson's "legend captions" doesn't provide enough information to document the changes he has made over the original version, and contains much irrelevant and even incomprehensible data. Matthews does not give enough information about data sources, and particularly the method for assigning seasonal albedo values.
19) Documentation written by original authors to fit a prescribed format, then edited by you, might be more useful. The authors of the test database documentation did not seem very knowledgeable of the development of the sets, nor of who their true authors were - citing Ray Milleman as author of Olson's Database.
20) Now it is.
21) In general, I found the documentation very helpful and a delightful collection of references. However, there were certain aspects of documentation I found confusing.
Although I liked the Global Vegetation Index User's Guide as a reference, it's inclusion with the NDVI dataset caused a bit of confusion. At first, I wasn't sure whether or not calibrated NDVI was included as part of the dataset. Then, I wondered why the calibrated dataset was not included with the dataset. If there were controversies in determining the calibrated NDVI, they were not apparent from the documentation.
Why was the Leemans and Cramer data considered "Monthly" data when it was averaged monthly? To separate it from Legates and Willmott? Why not label the sections Leemans and Cramer data and Legates and Willmott data?
22) Unfortunately, I do not have time to edit the documentation for this review. I think the current format is reasonable if one prefers a relatively rapid production of this database. Ideally, of course, more standardized and less individual information would be most useful.
23) Need better installation procedures. The documentation that was provided did not work on our PC. (Append command doesn't work.) Need more information about files so we don't have to use other software on PC's.
24) 1. I truly did not enjoy putting the documentation together.
2. The title of the database should be changed. It should read something like "Global Change Databases" or "Earth Science Databases"
25) Doc files look well organized and uniform.
26) Yes - very good.
27) N/A
28) The documentation for data transfer to GIS packages was inadequate with the exception of the IDRISI package. If data conversion software to convert the data to common spatial data formats (see 1c) had been provided extensive documentation on the details of the data format would have been unnecessary.
29) The documentation is rather complete.
30) Documentation on data - itself appears good, more help with implementation would be useful.
31) Yes, the documentation is quite good except for the cases below where the packing scheme was unclear.
32) Were the instructions clear and accurate for reading and displaying the contents of the database?
The Introduction section is clearly written, but Appendix A (IDRISI Implementation) leaves much to be desired if (as I attempted to do) the subset of IDRISI included with the database is to be the primary means of browsing through the data. Much of the problem appears to relate to the interface between IDRISI and DOS (and how that relationship is described in Appendix A). In particular (pages referenced being to Appendix A):
(a) On page 1, there appears to be a typographical error ("Append /E" should actually read "APPEND \E" - that is, a backslash should be used). [Editor's note: This is actually incorrect. The proper syntax is forward slash, i.e. APPEND /e -- please refer to the User's Guide or your DOS manual]
(b) Near the top of page 3, "ALL." should be "ALL" - that is, the period should be deleted.
(c) The use of ACCESS and APPEND should be more clearly described. In my installation, even though the APPEND path appeared proper, based on the response to entering APPEND at the DOS prompt, the environment shown in response to entering ACCESS did not include the APPEND path. Nevertheless, IDRISI appeared to be able to access the image files (but, for some reason, not the Micro World Data Bank II vector files) when "color a" was selected, although not when "list" was selected. In the latter case, IDRISI responded that the file was not found in the default directory.
(d) For users depending on the provided subset of IDRISI, it would be useful to provide some minimal level of documentation of IDRISI. It took some amount of trial-and-error to determine which commands were needed to display a map of each dataset.
(e) I used IDRISI to browse through each of the 11 chapters (looking at the first dataset in each chapter, if it contained more than one - e.g., GMVEG in Chapter 6). There was a problem, possibly specific to the personal computer system used, with FAOSOIL (but not GFAOSOIL) - the file appears to be have been read, but only the title and legend, and not the mapped data, appeared on the screen - perhaps the file was too big. If this is a widespread problem, a warning about system requirements would be useful. And, as mentioned in (c), the Micro World Data Bank II vector files were not read, perhaps a result of the way that IDRISI was used; as mentioned in (d), more instruction in IDRISI might have helped.
33) N/A
34) N/A
35) N/A
36) N/A
37) Needs some improvements, especially on origins of the data,
quality checking, etc. The user beware caveat is too easily lost
in the rush to have results.
VIII. DO YOU THINK THAT WE HAVE ADEQUATELY ADDRESSED THE ISSUES
OF DATA QUALITY? NOTE THAT DATA QUALITY ALSO REFERS TO DATA DOCUMENTATION,
ALLOWING THE USER TO JUDGE THE FITNESS OF THE DATA FOR A PARTICULAR
USE.
1) N/A
2) A broad statement spelling out the nature of the data and the responsibility of the user would help at the very beginning.
3) No, I don't think that you have adequately addressed the issues of data-quality. You have compiled a large series of very useful datasets, reformatted them and presented them in a unified format. I find this already a major achievement. Currently, the CD-ROM contains only the datasets and some of their documenting papers. No quality assessment is done at all. I would like that the contributors not only write a short summary describing the data (see also point 7), but also a similar note on the scope and limitations of their data. I foresee that many quality reviews of the individual datasets will be published in the scientific literature when the data becomes generally more available. If no notes are included on the limitations of the datasets, negative reviews of individual datasets could start rumors that the global datasets are trivial. The addition of a critical note could remove such threats.
4) Yes, unless your user community is very limited and well defined its difficult to fairly assess data quality. Its more important to provide data availability. Let the end user decide for himself what's best.
5) It's really the responsibility of users to read the original papers and judge the data quality for themselves.
6) The uncertainties and assumptions associated with each dataset are not clear.
7) Based on my initial review, I think you have. As noted below, the documentation in IDRISI and World Data Bank leave something to be desired.
8) In most cases, yes. In the case of NDVI, however, there could be more documentation on the potential uses, calibration to plant production, and limitations.
9) I think you have done as well as you could. However, interpretation from the maps, I believe, could lead to significant errors.
10) Data quality is always difficult to assess: particularly where datasets have been patched together to provide the global coverage. The ETOPO5 terrain data does not register perfectly with other mapping (in places). Unfortunately I did not have time to try a superimposition on ETOPO5 and the DCW prototype vectors. I imagine there might well be some problems.
11) Yes
12) Except as noted in 7), yes.
13) Again, a detailed reply is not possible since all our time was spent trying to access the data. I am somewhat skeptical, however, of the quality of the data contents if the same people prepared it as prepared the operational documentation.
14) It is difficult for those who compile different kinds of data to adequately address data quality issues. On the other hand, it might be possible to establish certain minimum criteria for the dataset overall and to indicate whether or not each author has verified the quality of his data to the extent of those criteria. Of course, each user of the data must also realize that the data are provided for use as is and that quality assurance is to some extent the responsibility of the user.
15) Data quality looks to be excellent based on the amount of time and effort to refine and document the process that was used to compile it.
16) The error analysis should be included for all datasets if possible. The raw data should be available if the user would like to conduct his/her own error analysis.
17) I actually see little evidence that the issue of data quality has been addressed at all. In the integrated bathymetry/elevation dataset you refer to a mis-registration without following up on the problem. I do think you did a good job in explaining some of the quality control that went into converting that data into a format usable in the database (but there are specific concerns - see comments above). I actually think that, in general, there is very poor overall documentation as to whether a data is fit for a given use (see comments above). Given the EPA's concern over quality assurance issues (and my own agency's as well), I think that more time needs to be spent covering the quality and usefulness of the data.
18) For now you can only use datasets that exist; but I would like to see some incentive for the providers of vegetation datasets to be more explicit about their classification procedures.
19) I had no problem but then, I only used the databases for which I had personal knowledge.
20) Yes, of course.
21) Other than noted above, yes.
22) cf my answer to question 3a). I think, more could be done, and specifically, I think the problem of resampling should be addressed more. One solution would be to always include the original data in some practical format.
23) Yes, dataset description in the .doc files and in the manual provide enough information for determining whether or not data is adequate for our use. It isn't yet. We are really only going to use vegetation because we need it for January, 1988.
24) Yes
25) Don't know.
26) Yes
27) N/A
28) Did not review data content (see 3).
29) It is difficult to answer the questions because we do not have enough time to analyze the data quality and completeness to a required extent. However, we promise to send to your address all our comments which may arise in the course of our work with the data.
30) For the FNOC data there appeared to be little documentation beyond basic description of data format. I couldn't find any discussion of data quality or much on historical development.
31) Again quite good; but more complete documentation/citations should be provided for all datasets.
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) " [see question #7]
IX. PLEASE LIST THE DIFFERENT TYPES OF APPLICATIONS THAT YOU
OR YOUR FACILITY ARE INVOLVED WITH.
1) N/A
2) GIS, hydrology, vegetation modeling, atmospheric density, atmosphere/biosphere interaction, global environmental management.
3) My department is involved in research on different aspects of the global change. There are two main fields. The first is the carbon cycle modeling with emphasis on global and regional emissions, global land use patterns, carbon fluxes, and population growth. The second field is modeling the impacts of global climate change on eco- and agro-systems. We use global climate datasets, climate vegetation relations and vegetation datasets for different applications.
4) N/A
5) I'm interested in predicting where and how well different tree species and provenances will grow under present and future environmental conditions. (see papers enclosed).
6) Measurement and modeling trace gas emissions including:
- methane emissions from cattle, rice, termites, natural gas distribution systems
- non-methane hydrocarbon emissions from vegetation
- NO and N2O emissions from soils
- reduced sulfur gas emissions from soils and vegetation
7) Developing and distributing basic data to support spatial research and generation of maps and map based products.
8) Modeling of tropical C and N cycles and trace gas fluxes from soils. Global respiration modeling.
9) We are involved in monitoring mitigation and adaptation of gas exchange between the atmosphere and biosphere, ecological dynamics, disturbance phenomena such as fire and insects, and human interactions.
10) Our work has been mostly local scale, but we are hoping to move into environmental global models and modeling in a general way with our final year undergraduate and Master's course students. I have started assembling pieces that can be made into a course that would serve as an introduction to the possibilities of using regional and global scale data. Your CD-ROM, and its hoped for successors are invaluable in this respect. In many ways more useful than the DCW data as it is presently easier and faster to handle.
11) World Climate Research Program's Global Energy & Water Cycle Experiment (GEWEX); Hydrologic Forecast Models.
12) In my present position I am assembling a database of narrow endemic vascular plants and vertebrates of the Pacific Northwest, their distribution and habitats, and collecting information necessary to assess the potential impacts of global warming on these organisms.
13) Global Climate
- scenario development
- GCM scenario analysis
- integrated regional modeling
- climate impact assessment
- EMAP (Forests and Integration and Assessment)
- ecosystem monitoring
- ozone
- wet deposition
- dry deposition
- air chemistry monitoring (wet and dry deposition)
- background studies
- baseline definition
- within ecological resource assessment
- trends identification and description
- associative studies
- across ecological resource assessment
- interpretive studies
14) ......basic and applied research in the atmospheric and planetary sciences. [A] staff of about 60 scientists and support personnel work in the fields of atmospheric chemistry, meteorology, climatology, physical oceanography, remote sensing, numerical weather prediction and planetary physics.
15) Our applications at this time are working and validating tree growth and stand level process models using data we will collect experimentally and that available from other sources such as Global Climate database.
16) ....analysis of large and small scale climate research. We have been involved lately in the determination of rainfall from satellite datasets and the verification of existing satellite-based rainfall algorithms.
17) Researchers at our facility are involved with a great number of topics. My own work deals with the influence on climate on the distribution and abundance of North American birds. I am modeling how birds respond to climatic influences and will then try and project how a given species' range may change under various global warming scenarios. Other work done here includes studies on population changes in neotropical migrants, impacts of global warming on wetlands in the Prairie Pothole region, impacts of grazing and land use practices on distributions and abundances of birds in the northern Great Plains region, and many issues concerning waterfowl population trends and biology. Many of the datasets in this database would be very useful to work going on here at the center. While a North American database would be even more useful, this data can be sub-setted and provides a very good start. I am particularly looking forward to using this dataset in GRASS once we get our new GIS lab set up.
18) - Development of global biome models
- Application of global biome models to paleoclimatology and paleoecology, including data model comparisons
- Coupling biome models to atmospheric general circulation models and analysis of implications for ecosystems
- Development of vegetation dynamics models
- Analysis of vegetation sensitivity to global warming scenarios and CO2 change.
19) 1. Research in developing global biosphere models ....
2) Research in modeling tree migration in upper Great Lakes and Canada.
3) Graduate education in global change (use in seminars, GIS courses, graduate research).
20) N/A
21) I'm not sure I understand the question, but here's an attempt to answer it:
.....basic field research programs studying ecosystem processes (including trace gas fluxes) in:
a) lakes
b) streams
c) salt marshes
d) oceans and inlets
e) tundra systems
f) temperate forests
g) tropical forests
In support of these activities, the Center has been developing several models to examine ecosystem processes and the effects of various factors on those processes. In addition, the center is developing global- to continental-scale ecosystem models to describe the effects of climate and other factors on ecosystem processes.
[Cover Letter] Our group is interested in trying to link the results of field studies to the effects of global change in ecosystem processes (e.g., NPP, net N mineralization, trace gas fluxes) at continental to global scales. We are developing spatial models to provide this link.....
22) - global change impact on terrestrial ecosystem scenarios, using ecosystem models
- education of geography students on global ecology, database management and GIS
23) 1. Climatic research.
2. Intraseasonal climate variations research.
24) 1) Carbon cycle modeling
2) Agricultures contribution to Global Change
3) Potential C sequestration in agriculture
4) Earth albedo
25) Global Climate Change
US/North American agriculture carbon cycle interactions.
26) CD-ROM general reference/education atlases
27) N/A
28) 1. Atmospheric Research
2. Visualization for scientific research
3. Exposure Modeling
4. Risk assessment
29) The database has been used in the course of the global climate change analysis and the investigation of their connection with regional anomalies.
30) Earth Remote Sensing (RS)
RS product processing
RS product calibration and validation
31) 1) First ISLSCP Field Experiment:FIFE
- Association between satellite data and terrain variables
- Estimation of surface biophysical variables from satellite data
2) California Biodiversity Project
- vegetation mapping and ecological modeling of habitat types for entire state of California
- GAP analysis to identify ecological preserves based upon biodiversity criteria
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Crop-yield modeling -- daily + monthly water balance + agroclimatic index; point + spatial scale
Resource/population/land use models
World food trade
Landscape vegetation zones
Energy demand
Greenhouse gas sinks
(mostly all corrected to climate impact assessment)
X. WHAT APPLICATIONS DO YOU SEE THE GLOBAL DATABASE BEING USED
FOR?
1) N/A
2) see (9) and educational purposes in training workshops.
3) Education: With these global databases it becomes more easy to present many kinds of global scale patterns, to link and evaluate them. The CD-ROM could in time become more useful for education of our physical global environment then an atlas.
Research: The database could be used to describe species distributions in terms of climatological parameters. Some research has been done already by no there becomes data available for continental and global applications. If the CD-ROM is available throughout the world many new and unforeseen application will surely be developed.
4) N/A
5) Ecological analyses
Agricultural/Forestry analyses
Under both present and future conditions.
6) Global biogenic emissions, which are a function of soils, vegetation type, temperature, light intensity, biomass density, and other factors.
7) Clearly, environment, weather, and geography research and education comes to mind. The topography data will probably be used in a wide variety of research, education, and production work. You might consider asking your customers to let you know what they are doing with the data.
8) Spatial ecological modeling and biogeochemistry changes in response to climate change.
9) I think it should be useful for reference and base-line data from which to build more definitive sectoral data.
10) Global/regional modeling of atmosphere, environmental change. Hazard assessment, monitoring, regional assessment of reinsurance risks (given geological overlays etc).
11) Global Hydrologic Modeling and Improved Parameterization of GCM's.
12) Regional and large scale studies of interaction between physical environment and natural systems, and/or physical environment and human-dominated systems seen to use the most appropriate uses of the global database. These studies would include generalized descriptions of individual regions. Regions with similar environments in different locations or regions with contrasting environments could be usefully compared.
13) For [our] activities, the database's primary use would be in the development of background products and the definition of baseline conditions against which ecosystem status and trends assessments could be compared. For climate studies, the database would be used as a baseline against which GCM baseline output would be compared. It could also support ongoing research into techniques for the derivation of more detailed regional climate change scenarios.
14) In addition to using geo-political boundary and elevation data to plot maps for use in regional data studies, [our program] scientists have expressed interest in using a global database such as the GED; (1) to verify surface vegetation types for satellite-based cloud studies and (2) to obtain surface boundary condition data for use with numerical weather prediction and general circulation models.
15) We envision some kind of spatial representation of the Forest Ecosystem with separate layers to include different plant and tree species, growth through time based on interactive input, and linkage to expert systems as they develop for different types of modeling environments.
16) They are directly applicable to construction and verification of climate models and the assessment of climate and climate changes. The packing of several dataset on one optical disk will allow easy intercomparison of the datasets.
17) At present I see the global database being used primarily for biogeographic and terrestrial ecology questions. This is not necessarily all that bad. Before we can understand what impacts global climate change may have on plants and animals we need to have an idea of how they respond to "normal" conditions. Most ecological and biogeographic studies have had a very limited geographic coverage. This has largely been due to, I think, the lack of available data and lack of understanding of the scales involved. The global database goes a long way toward resolving the issue of data availability. Scale issues are important and not so easily solved but that is not the job of the database. I reiterate the point that I think that better data extraction routines need to be written that will allow users with no availability to sophisticated GIS to be able to access the data. Without the data extraction routines I think you might risk being perceived as collection of pretty pictures.
18) All kinds of work involved with global ecosystems modeling, global hydrological modeling and biosphere - atmosphere modeling.
19) see 9 above
20) The decision of the forecast problems. The different model creation (e.g., models for CO2 increase and others).
21) The global database will primarily be used as input datasets for the modeling efforts. However, data may also be used to help put together summaries for publications related to the field research program.
22) The ones listed under 9) (we are a geography department with all standard activities usually found at such a department; the ones I mentioned are the ones of relevance to this product).
23) 1. Comparisons with GCM's
2. Comparison with analyses like winds and moisture fields.
24) At least everything I have listed in question 9.
25) Climate Modeling. Once equivalent versions of the same data categories have been gathered for a long enough period of time, historical trend analysis - prediction will be possible.
26) I can't speak to any but our own, but I am sure you will be surprised by unexpected uses.
27) N/A
28) N/A
29) The database can be applied to research in hydrometeorology, ecology, for environmental monitoring, etc.
30) As a basis for RS (remote sensing) data to be developed and understood. As understanding of a particular dataset develops it may itself become part of the global database - once proven adequate to agreed standards. In this way, the global database may become an international standard.
31) - Exploratory analysis of global scale vegetation - climate interactions
- Continental phenology studies
- Change detection of large scale practices of deforestation
- Analysis of time-space properties of satellite data.
32) N/A
33) N/A
34) N/A
35) Morphometric analysis of 5 and 10 minute DEM data using classification system of Hammond.
36) Evaluation of Climate GIS Databases for Identification of Crop Environments: case study for Northern Latin America.
37) Global and regional Agro-Economic Zone project -- [as described
elsewhere]
XI. WHAT PROBLEMS DO YOU FORESEE IN MAKING THIS DATABASE AVAILABLE
FOR EDUCATIONAL PURPOSES AT VARIOUS LEVELS? THIS MAY INCLUDE HIGH
SCHOOL THROUGH COLLEGE, PUBLIC EDUCATION, TRAINING PROGRAMS FOR
SCIENTISTS AND MANAGERS.
1) N/A
2) (1) Training the educators.
(2) Obtaining appropriate hardware.
3) One of the problems that I foresee is the scale of the dataset. Most of the data sets are on a 0.5 degree or even coarser grid. For educational purposes users will not be fully aware of the limitations of such medium scale resolution and use the data for only small regions, like their own backyard, and find discrepancies with the observed patterns there. I think that a thorough discussion on different scales, scope of applications, etc., should be included as a guidance for the potential user. Unfortunately, I do not know of a paper that addresses this issue.
4) Maintaining a level of documentation that would serve both communities. The educational community would prefer a canned tutor session or sample session to get by some of the technical expertise needed otherwise.
5) None foreseen.
6) N/A
7) As with any new product, simply getting the word out to the intended audience is the first problem. The second will be with the schools getting, or allocating, the money to purchase the database, IDRISI, and the CD-ROM drive if they don't have one. Getting the product to the schools will only be the first part of the problem. If the rest of the IDRISI documentation is like what is in the manual I received for this review the bigger problem will be in using the data - especially at the high school level. For the school environment I would strongly advise that the IDRISI documentation be considerably expanded and include a tutorial. The value of the data will be greater if it is supported by software which is geared to that level. Based on the flyer on IDRISI I received from Clark University it appears to me that many of the IDRISI capabilities are quite advanced for the high school level. For this level some other less technical program might be advisable. In this respect, the IDRISI documentation is the low point in the package I received.
8) Integration with dynamic models.
Lack of data at various spatial scales (i.e. 60 minute)
9) The main danger is making errors in interpreting the data, and extrapolating the information beyond the applicable scale.
10) No problems at all! It would be an idea to develop a good interactive teaching pack to help the uninitiated (academics as well as students) operate prepackaged courses, and to include a genuine modeling (in the Spans sense) capability, possibly based on IDRISI, and certainly based on nothing more elegant than a PC386 with Super VGA. There is nothing more irritating in the educational environment than to discover that your 20 PCs don't have the vital piece of specialist hardware required. Packages (and people) get thrown out of the window when that happens.
11) N/A
12) As the database becomes more widely distributed, I would foresee problems both of accessibility and (over-) interpretation. I do not think CD-ROM readers are available uniformly; at lower educational levels the problems probably become worse. GIS expertise is probably equally limited in distribution. As currently distributed, problems in producing hardcopies of plots would also limit the accessibility of the information, I found the "plot" module of IDRISI necessary for combining raster information of interest with vector files from Micro World Data Bank II that provided geographic reference. I wasn't able to produce a hard copy of the plot though, and I imagine most users would lack the hardware to do so. At the same time, the farther the information gets from the compilers, the less appreciation there will be of inherent errors due to resolution of input data, resolution of distributed files, methods of interpolation, etc. Pointing out the usefulness and the limitations should be a major component of documentation and promotional material.
13) The biggest drawbacks at this point are the poor access documentation and the access speed. In its present form, the data transfer time is excessive (this refers strictly to the hardware aspects, not those associated with the documentation).
14) N/A
15) No problems in making the database available for educational purposes, only possible misuse of the data in not analyzed within context - extrapolations, statistics, unscientific procedures, conclusions.
16) There should be no problems. However, they may be of limited use due to the lack of sufficient hardware and software and the expertise to use them. Training programs will be necessary.
17) 11 and 12) I have lumped these two categories because many of the comments are applicable to both. In some respects it seems that you need two different copies of the data acquisition and view software for the two different groups. I think that the database itself is useful for many different levels of use, from high school through research. I would suggest the following however:
1) Put the data acquisition software on floppy diskette rather that on CD-ROM. The benefit to putting the software on floppy is that you can have two different versions of the software to access the same data (see below).
2) Create two different versions of the data acquisition software. It would seem to me that high school, college classes, public education, etc. would mostly be interested in actually viewing the data. I would suggest that you provide software that would allow for easy viewing/zoom viewing of the data, possibly the ability to string several images together into a slide show, and the ability to export the color images to another format (i.e. a screen capture utility that would export the image as a .PCX (or other) image). The export of the color images would allow for the data to be used in many different novel ways. The second version of the software would be similar to that distributed now with some of the recommendations that I and I'm sure others have made.
The problems I see in distributing this database are largely financial. Most CD-ROMs are overpriced in comparison to the cost of manufacturing them. Even the $500-$600 charged by the NGDC for other datasets is probably excessive when compared to the average charges levied by EROS for CDs of its data (average cost $35). The cost issue is especially important if you wish to attract users in the education and public sectors. Another problem might be the acquisition of rights to publish copyrighted material (which I assume much of the supplemental material that you included to be). I think the inclusion of this previously released material is very important and should be pursued.
18) The database as a whole is not really suitable for this purpose because many of the datasets are too experimental, the documentation differs strongly among datasets, etc.
19) It probably should include the GIS (IDRISI), at least as an option, when others obtain it. Still, its not a terribly user-friendly system for public schools - a good PASCAL programmer could build a windows interface which is particularly useful in such complex applications which are only occasionally used by operators.
20) N/A
21) Technically, I see no problem in making the database available for educational purposes. I think most institutions can afford the necessary equipment, if they don't already own it.
The problems I foresee in making this database available to educational institutions is the extra time and effort it will take for people to learn a GIS in order to access the data. In addition, I think that once people have developed GIS skills to access the data, they will be frustrated by the lack in the availability of spatial data necessary to carry out their specific analyses. But, you've got a start somewhere and I think making this CD-ROM available is a good start.
22) Frankly, I don't foresee great problems. IDRISI is in the process of getting more and more established as a general low-cost GIS, and any IDRISI-user with a PC and a CD-ROM drive can access this database.
23) For educational use we should have all fields at all resolutions.
24) Audiences other than "Highly Technical" will likely have no use for this data nor will they appreciate its value. For those audiences, the datasets must be "Put into the appropriate contexts and perspectives", i.e., a sort of interactive slide show of global change information.
25) Simply need to ship appropriate access software (or scripts/macros) for large, in-place applications, such as for ARC/INFO). Then let the users integrate it into their GISs.
26) Data may be too raw for non computer-literate to use.
27) N/A
28) For educational purposes the IDRISI software may be appropriate depending on the audience. I do not believe the function of the databases on CD should be educational. Maybe two sets of CD's (one educational and one for research) would be more appropriate. I believe in the effort to make the data useful to "all" possible users the authors of this CD have made the data inaccessible for most GIS users without special programming skills.
29) N/A
30) I see no insurmountable technical problems. However, the data are open to misinterpretation by those remote from the point of construction.
31) 1) Technology: many public education facilities cannot
afford the hardware to extract and display this database.
2) Expertise: appropriate use of these data requires high
level of training regarding data limitations, sources, and appropriate uses.
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Equipment requirements are relatively high (for UK Academics
and particularly for schools). Will need "packaging"
-- original slide show diskette was better for a quick look --
some sort of hyper-text menuing/interpretation system but these
could be added or tailored by individual users. What about developing
a menuing prototype that we can alter? We give 100s of demonstrations
a year, including IDRISI, project results, PCX graphs, screen
dumps, etc.
XII. WHAT PROBLEMS DO YOU FORESEE IN MAKING THIS DATABASE AVAILABLE
TO VARIOUS SCIENTIFIC GROUPS? THIS MAY INCLUDE GLOBAL CHANGE RESEARCHERS,
RESOURCE MANAGERS, AND A VARIETY OF SCIENTIFIC GROUPS IN VARIOUS
FIELDS.
1) N/A
2) - Advertise its availability in EOS regularly.
- Make it available over the network.
- Provide workshops around using the data.
3) The quality of the dataset becomes more important for research issues. If no statements on the quality and limitations are included, the research community will make their own judgement, often based on unrealistic expectations and an unfortunate misunderstanding of the evolution of many of the datasets.
Hopefully many of the contributors of current datasets will present new, better versions in the near future. In the creation of these new versions, an important task for your group will be to guide the creation of the proper, comprehensive and compatible documentation describing all aspects of the datasets. Only if such documentation is available the dataset should be included in the global compilation.
4) Getting to a common level in terminology and to avoid compromise that hurt the integrity of the data or make it suspect to the data purists.
5) People will like it so much, they'll want even more data.
6) The main problem I would foresee from the wide dissemination of these data is that it could be used without a good understanding of the uncertainties and assumptions associated with each dataset.
7) I suspect that the scientific and technical community will tend to pick up on the availability of this database faster than the academic community. Cost should also be much less of a problem for these groups. In both cases it will simply take time to get the word out.
8) Lack of recognition of the limitations of data sources.
9) I foresee the same limitations as with students, but with less probability for misinterpretation of data.
10) No mention has been made about cost... I would hope that the final product will be as cheap as possible! For teaching purposes in my estimation a price over $1000 would kill the project dead, $100 would mean bonanza sales. Research and business would not have the same financial constraints. Obviously only 1 copy is needed per establishment, but there may be problems with licenses on the basic GIS (very necessarily) distributed with the GED.
11) Only the cost which can't be very high.
12) Probably the largest problems here are hardware software, and personnel concerns pertaining to data accessibility. CD-ROM readers may not be easily accessible to many scientists and managers. Using major libraries (which presumably do have CD-ROM readers) [for] data distribution might help. The large number of GIS systems in use, and the unavailability of any GIS to some individuals also may constrain accessibility. Finally, in agencies or institutions with centralized GIS functions, my experience is that personnel are already quite busy. Adding mediating use of this database to their responsibilities may not be feasible.
13) I do not think that IDRISI is the way to go. Also, unless access, transfer and display speed on the PC can be improved, I think this group ... would really rather opt for workstation (this system could not be made to work on a DEC workstation either), mini (VAX) or mainframe versions of the system. Finally, the IDRISI menus do not contain an export function. This is necessary not only for GIS file conversion, but for data use in non-GIS settings. At the very least, a standard ASCII file format should be offered as an output option.
14) Storage of files in IBM or DEC readable binary format might discourage individuals or groups that do not use IBM or DEC machines from using the data. We, for example, do not have the proper equipment to use the special GIS CD-ROM access software that was provided on floppy diskette for this review.
15) Similar to above - using data out of context or disregarding extrapolations or limitations in collection and/or compilation. Misunderstanding of these same. Example: Spotted Owl habitat data in the Pacific Northwest released to public and used out of context to influence legislation.
16) I do not foresee any major problems. A wide distribution will allow more interdisciplinary research which is vital for the climate change problem.
17) [see Question #11]
18) Even in this case, some attempt at a more standardized documentation format would help.
19) Probably, the same generic problem and solution as in 11, above.
20) N/A
21) I think global change researchers and resource managers need real time data to carry out their respective analyses appropriately so that the data on the CD may be of limited use to these groups. In designing our model..., we recognized the lack of spatial data needed to develop a complex model so we designed [the model] to be a rather simplistic ecosystem model. Hopefully, as more datasets become available, more complex models can be developed.
As for other scientific groups, I think the same problems listed above for educational purposes applies here. Time is a precious commodity so the benefits from the available dataset need to overcome the costs of learning the GIS needed to access it.
22) Users have to figure out how to get the data out from IDRISI and into their application.
Also, they have to be made aware of the unavoidable limitations (and their spatial pattern) of such data. The really big danger is that users assume this kind of data to be THE TRUTH. I can, however, not give much advice on how this problem can be overcome.
23) Hardly any scientist uses a PC. There needs to be a more generic database that a wide variety of platforms can access. Unix and Fortran on workstations are the most widely used systems of scientists.
24) The acquisition costs must be "low to free". Many scientists will be unwilling to pay for data that is available for free from other sources.
25) Ditto question XI, #25).
26) None
27) N/A
28) Since intense scientific use will undoubtedly uncover QA/QC problems with the data, managing the updates to future revisions will be a challenge. How the updates are selected and applied to improve future releases of the CD will be a major task. In addition, data ownership issues (political, commercial,...) will also have to be overcome if more data is to be made available for distribution.
29) N/A
30) None - except as above!
31) The major issue here is using simple, accessible data format with good documentation. As long as these issues are given continual high priority I foresee few problems.
32) N/A
33) N/A
34) N/A
35) N/A
36) N/A
37) Data quality issues and desire to go back to original sources to avoid (a) other's errors, (b) issues of copyright and publication credit.
What about an e-mail network on global databases and applications?
A quarterly electronic newsletter would help keep in touch!!