IUFRO conference on Growth studies in tropical moist forests in Africa, Kumasi, 12-14 Nov 1996

1. Abstract

TROPIS, the Tree Growth and Permanent Plot Information System, will contain five elements:

• A network of people willing to share permanent plot data and tree growth information (serviced by newsletters and information sources hosted at http://www.cgiar.org/cifor or available from CIFOR),
• An index to, or metadatabase of, people and institutions holding permanent plot data,
• A database management system (MIRA) to assist more efficient data management,
• A system (WORLD) to identify comparable sites in other regions, allowing data from elsewhere to be used when no local growth information exists, and
• An inference system (PLANTGRO) to allow growth estimates to be made in the absence of empirical data.

Most of these components are still under development, but the first two elements are operational. The index or metadatabase currently contains references to 10,000 plots with 2000 species contributed by 100 collaborators, and is growing at about 1000 plots per month. Searches of the database are welcomed, and may be directed to the author.

2. Introduction

TROPIS is the acronym for the Tree Growth and Permanent Plot Information System sponsored by CIFOR, the Center for International Forestry Research, to promote more effective use of existing data and knowledge about tree growth. Several recent reviews report a paucity of long-term studies in terrestrial ecology (e.g., Strayer et al. 1986, Tilman 1989); this presumably relates more to the availability of data from long term permanent plots, rather than the existence of such studies. TROPIS attempts to redress this situation by drawing attention to existing studies. TROPIS is concerned primarily with information about permanent plots and tree growth in both planted and natural forests throughout the world. It has five components:

• a network of people willing to share permanent plot data and tree growth information;
• an index (metadatabase) to people and institutions with permanent plots;
• a database management system to promote more efficient data management;
• a way to find comparable sites elsewhere, so that observations can be supplemented or contrasted with other data; and
• an inference system to allow growth estimates to be made in the absence of empirical data.

3. The network

TROPIS is about people, and about information. So unless they request otherwise, all contributors and clients of TROPIS are placed on the mailing list of TROPIS-Update, a twice-a-year information sheet informing of recent developments and progress. At present, TROPIS-Update goes to 200 people by email, and to 200 people by regular mail. Others are also welcome to subscribe, and may do so by sending an email to listserv@cgnet.com with the message subscribe tropis or by contacting the author.

TROPIS also provides information via the internet, at http://www.cgiar.org/cifor in the USA. Here, in addition to the latest information on TROPIS, how to contribute, and how to search TROPIS, you will find pointers to other sources of long-term permanent plot data. Although these other sources may serve different objectives and have different scales, they are consistent with the TROPIS objective to make better use of existing data. If you are aware of monitoring efforts not documented in the TROPIS internet pages, please bring them to the attention of the author.

4. The index

The core of TROPIS is the index to (or metadatabase of) people and their plots, maintained in a relational database. The database is designed to fulfill two primary needs:

• to provide for efficient cross-checking, error-checking and updating,
• to facilitate searches for plots matching a wide range of specified criteria, including, but not limited to location, forest type, taxa, plot area, measurement history, etc.

The database structure is outlined in Figure 1, and in the data entry form in the appendix. the database is essentially hierarchical: the key element of the database is the informant. Each informant may contribute information on many plot series, each of which has consistent objectives. In turn, each series may comprise many plots, each of which may have a different location, a different size, etc. And each plot may contain many species. A series may be a thinning or spacing experiment, some species or provenance trials, a continuous forest inventory system, or any other aggregation of plots convenient to the informant. Plots need not be current, and abandoned plots may be included provided that the location is known and the plot data remain accessible. In addition to details of the informant, we try to record details of additional contact people associated with plots, to maintain continuity when people transfer or retire. Thus the relational structure revealed in Figure 1 may appear complex, but ensures data integrity.

At present, searches are possible only via mail, fax or email requests to the TROPIS-coordinator at CIFOR, but it is anticipated that self-service on-line searching will be made available next year (assisted searches will continue to be available for those without Internet access).Clients may search for plots with specified taxa, locations (latitude/longitude or place name), silvicultural treatment, or other specified criteria and combinations. Some requests previously fulfilled include searches for

• plots with particular species and/or locations (regions, latitudes, elevations, etc.);
• plots planted with two species and a range of spacing and thinning treatments; and
• plots in logged natural forest with several remeasures spanning at least 40 years.

The main outcome of such searches is a list of people to contact, with details of the nature and amount of relevant data held. A catalogue of past searches is also maintained, so that clients with similar requests can be advised of their common interests.

TROPIS currently contains references to over 10,000 plots with over 2,000 species contributed by 100 individuals in all parts of the world.

5. Database management

Several sources, including some of the early contributions to TROPIS, indicate that many researchers have some difficulty in compiling field data into an efficient database. Informal surveys of contributors who store data as paper or word-processor files revealed difficulties with basic technical aspects of database design, often rather similar in nature. TROPIS attempts to eliminate some of this unnecessary duplication by providing a standard database system free to any contributor who requests it. Such standard database systems have been proposed before, often with limited success, but one such system, MIRA (Ugalde 1988, 1989), has been used extensively in Latin America for several years, and appears to meet the basic requirements of many TROPIS participants. Thus CIFOR has sponsored the development of a new version of MIRA, based on a standard platform (MS-Foxpro) and made multi-lingual (English, Spanish, and French, plus the ability to customize to a fourth language). The prototype is now undergoing testing, and the package is expected to be available next year.

6. Comparable sites

Researchers studying tree growth are often handicapped by the paucity of data, or by the absence of independent data to corroborate their findings. Tree ring analysis is not always possible, and in such cases, growth data must be obtained from direct measurement. Reliable growth estimates require permanent plots that have been remeasured regularly for long periods, and these are not always available. However, there are many plots world-wide, and some of these may be used if an objective basis such as homoclime analysis can be used to select comparable growing conditions. Such analyses are commonly undertaken to assist species and provenance selection (e.g.,. Booth 1990a, 1991), but the issue of identifying comparable plots is analogous. Thus TROPIS will include a new version of WORLD (Booth 1990b) to enable such comparisons to be completed efficiently. This work is still in progress, but will be available by mail, fax, email and on-line when completed.

7. Objective inferences

Homoclime analyses are useful when data are available from comparable sites elsewhere, but this is not always the case. In some cases, despite judicious searches, no comparable data can be located, and yet it may still necessary to make some forecast about the suitability of a species on a given site. Despite this difficult situation, it may still be possible to provide a reasonable estimate, by making expert inferences from existing knowledge about the site and about the species under consideration. The PLANTGRO system (Hackett 1991) has been used with some success for agricultural crops, and is being enhanced so that inferences about tree growth can be made in the absence of empirical data. A preliminary version of PLANTGRO for trees is currently being tested. It too, will be made available by mail, fax, email and internet in due course.

8. How to participate

The objective of TROPIS is to help people learn more about trees and forests, and to help them manage these resources better. Anyone may contribute information on their permanent plots to the TROPIS system, provided that they agree in principle to share their data with others; subject of course, to a mutually satisfactory agreement between the data owner and the intending user. Conversely, anyone may use any of the five components of the system, provided that they agree to provide information on any permanent plots that they have, and respect the rights of other contributors.

Subscriptions to TROPIS-Update may be emailed directly to the listserver, or directed to the author, and anyone with internet access may browse the TROPIS internet pages. Information about permanent plots is welcomed from anyone with the appropriate authority, and may be submitted to the author using the form in the appendix, or using the form found in the TROPIS internet pages. At present, searches of the index must be directed via the author, but on-line searching should become available soon. The other components, MIRA, WORLD and PLANTGRO, are not yet available for general use, but will be made available soon. Their availability will be announced in TROPIS-Update.

9. Acknowledgements

I am indebted to the participants of the December 1995 workshop (Robert de Kock, Vitoon Luangviriyasaeng, Peter Muraya, Luis Ugalde, Tim Vercoe and Howard Wright) for their help in refining TROPIS concept into its present form.

10. References

Booth, T.H., 1991. Where in the world? New climatic analysis methods to assist species and provenance selection for trials. Unasylva 42(165):51-57.

Booth, T.H., 1990a. A climatic analysis method for expert systems assisting tree species introductions. Agroforestry Systems 10:33-45.

Booth, T.H., 1990b. Mapping regions climatically suitable for particular tree species at the global scale. Forest Ecology and Management 36:47-60.

Hackett, C., 1991. Mobilising environmental information about lesser-known plants: the value of two neglected levels of description. Agroforestry-Systems 14:131-143.

Strayer, D., J.S. Glitzenstein, C.G. Jones, J. Kolsa, G.E. Likens, M.J. McDonnell, G.G. Parker and S.T.A. Pickett, 1986. Long-term Ecological Studies: An illustrated account of their design, operation and importance to ecology. Institute of Ecosystem Studies Occasional Publication No 2. Millbrook NY.

Tilman, D., 1989. Ecological experiments: strengths and conceptual problems. In G.E. Likens (ed.) Long-term Studies in Ecology: Approaches and alternatives. Springer NY, pp. 136-157.

Ugalde-A., L., 1988. MIRA: un sistema de manejo de informacion sobre recursos arboreos (MIRA: a system for managing information on tree research). Centro Agronomico Tropical de Investigacion y Ensenanza (CATIE), Actividades en Turrialba 16(2-3):1-4.

Ugalde-A., L., 1989. The MIRA management information system for fuelwood and multi-purpose tree species research in tropical areas. Centro Agronomico Tropical de Investigacion y Ensenanza (CATIE), Serie Tecnica, Informe Tecnico No. 143:86-104.

Click here to see the structure of the TROPIS metadatabase.