FIFTEEN CENTURIES OF CHINESE ENVIRONMENTAL HISTORY: CREATING A RETROACTIVE DECISION-SUPPORT SYSTEM Robert M. Hartwell* CITAS and the University of Pennsylvania CSS94, June 2, 1994 During the late 6th century, the Chinese emperor Sui Wendi and his successor Yangdi mobilized an army of workers to join the two major West-East River systems of China by linking a chain of canals into the Grand or Imperial Canal, a North-South waterway that continues to support a heavy barge traffic in 1994. In 1012, an imperial edict mandated the transfer of early ripening rice seed from Fujian to other regions south of the Yangzi and ordered state agricultural extension agents to promote its use. During the 1070's, imperial advisors debated the means and effects of a plan to shift the channel of the Yellow River and then implemented a program designed to achieve this goal. In the mid-thirteenth century, Kubilai Khan reconstructed the northern portions of the Grand Canal and extended its length from its juncture with the Yellow River to a point east of modern Beijing. These decisions, and countless others, authored or anonymous, large-scale or incremental, were made without the assistance of computer-aided decision-support systems. They were, however, based on feasibility reports supported by quantitative data drawn from official surveys. The high cultural value the Chinese place on utilizing the experience of the past to inform the present and guide the future, and an information revolution created in the ninth century by the invention o printing, insured the preservation of vast portions of these voluminous record-keeping monuments to bureaucratic efficiency. Investigative reports of inter-agency committees, minutes of meetings of policy-making boards, feasibility memoranda, periodic census, fiscal and cadastral surveys, records of natural events, lists of local flora and fauna, and a myriad of other types of texts have left a paper trail, extant in printed and manuscript local histories, digests of legislation, encyclopedias, and government archives. These sources document an infinite number of decisions, the fund of knowledge and system of values on which they were based, and the immediate and multi-generational impact of these determinations. One example of scale will have to suffice. Zhang Peiyuan and his associates at the Institute of Geography of the Chinese Academy of Sciences (CAS) have assembled a precisely located and dated (day, month and year) dataset of nearly one million natural events (1-1949)--typhoons, floods, droughts, unseasonable heat, severe cold, excessive precipitation (rain or snow), bounty harvests, and famine crops (Zhang and Zuo, 1990; Zhang et. al. 1993). A high priority for the CITAS-CAS collaboration over the next five years will be to digitize the dis- aggregated data. Selective digitization of the Chinese historical textual databank will provide a basis for both micro- and macro-analyses of the human dimensions of environmental change. What were the considerations behind a project to alter the natural course of feeder streams, dam the outlet rivers, and impound the shore waters of a lake in order to expand acreage in cultivable fields and what were the immediate and long-term environmental consequences of this enterprise? Who decided, and on what basis, to impound coastal waters for solar salt production in a particular community and what were the short-term and permanent effects on the regional economy and ecology? On the one hand, documented examples of such specific cases of manipulation of the environment over the past fifteen centuries are nearly inexhaustible. On the other hand, these discrete projects took place within the context of large-scale regional transformations. During 1995-2000, CITAS proposes to continue to develop a cumulative electronic bank of historical and contemporary data in order to support research in three interrelated priority areas: 1) land use/land-cover change in China, 500-2000, 2) community manipulation of the environment and adaptation to its change, 500- 2000 and 3) the global impact of past, present and future environmental change in China. The land use/land-cover change database will be extensive and provide a millennium and one half county- and/or prefectural-level coverage of all of China for crucial variables--e.g. cropping system variables, harvests, yields, prices, climate, aggregate population, gross hydrology and major highways. The community studies database will also contain these variables. In addition, for the selected communities (counties), it will include additional micro-data. The concepts and methods of genealogical demography combined with data extracted from death records and lineage genealogies will yield geo-coded attribute tables consisting of such variables as marriage age, sex ratios, family size, occupational structure, mortality and/or longevity, epidemiological history, ethnicity, rural/urban distributions, migration, social structure, and kinship organization for sample groups and localities. For these same areas, local historical records provide the means to reconstruct 1) detailed changes in hydrology (e.g., irrigation networks and water conservancy projects), 2) refined alterations in the transportation grid (e.g. roads by class, bridges, fords, transport canals), 3) secular variations in political structure and decision-making variables, 4) proxy and instrumental measures of methane, CO2, sulphur, phosphorus and nitrogen emissions as well as secular variations in climate, and 5) detailed records of specific cases of human manipulation of the environment. As the temporal coverage draws closer to the present, these databases will tend to merge-- published statistics for the 1980s and 1990s will include much of the micro-data assigned to the sample communities for earlier periods. Finally, the disaggregated proxy and instrumental data on environmental variables will support research on the global implications of Chinese environmental history. In this way, the CITAS database will support 1) both macro- and micro-analysis of land use/land-cover change over fifteen centuries, 2) investigations into other aspects of the human dimensions of environmental change such as population and resources, political decision-making and alterations in local ecosystems, and the relationships between different configurations of social and cultural contexts and settlement patterns and the use of space, and 3) studies of the global impact of Chinese manipulation of the environment. CITAS priorities under the first Five Year Plan are 1) development of a database incorporating those sets of information required for research in the targeted areas and 2) creation of a decision-support delivery system that integrates RDMS with GIS, other graphics and statistical software, and computer simulation, game theoretical and fuzzy logic models. The goal of database development is to provide 1) a multi-layer contemporary and historical GIS, 2) complete, clean and integrated databases and 3) sets of attribute data in a form that can be used by social and natural scientists, global environmental change researchers and policy-makers regardless of their expertise in Chinese studies. The periodization of land-use/cover change in China over the past fifteen centuries illustrates the possibilities for comparative investigations into alternative environmental, cultural, institutional and socio-economic contexts for decisions made in response to, or impacting on, environmental change. Chinese land-use technology evolved over the past 1500 years through 1) innovations allowing for the expansion of multiple cropping by a) more intensive exploitation of fields already under cultivation and b) reclamation of hitherto sub-marginal lands and 2) the expansion of wet field paddy agriculture relative to dry farming (Perkins, 1969). Although this was a continuous process throughout all periods, there were at least six epochs of major regional transformations that occurred before the most recent period of accelerated change based on the introduction of modern science-based technology. 1. Completion of the Grand Canal in the late sixth century stimulated a period of rapid development in the Lower Yangzi physiographic region core (modern Shanghai municipal province and its contiguous hinterlands in Southern Jiangsu and Northern Zhejiang) and in other subregions situated along the canal course. This cycle of development abated around the middle of the eighth century. CITAS projects will reconstruct the adaptation of new cropping systems to the altered watersheds--the clearing of sub-marginal lands now brought under cultivation for the first time as well altered cropping systems in fields which had been exploited in the past. 2. Internal rebellion, foreign invasions and political instability stimulated a diaspora to the Lower, Middle and Upper Yangzi regions between ca. 880 and 1150. The agrarian technology which supported this increased population and sustained its continued growth was the adaptation of new varieties of rice seed as well as the extensive application of those developed centuries earlier, but in limited use (Kato, 1954; Ho, 1956). The result was a complete reversal of the relative proportion of the population located in the two major agrarian regions of China. Before 750, 2/3 of the Chinese lived in the dryland wheat farming areas of the North. By 1150, 2/3 of the inhabitants of the Song-Jin-XiXia empires farmed fields in the irrigated paddy agricultural areas of the South. This transformation was accompanied by at least one hundred per cent over-all growth in population from about 70 million in 750 to over 150 million by 1150. The new rice seeds included differential ranges of early ripening, late ripening, and drought resistant varieties which allowed for infinite incremental alterations in cropping systems adapted to maximize yields in different soil, drainage and climatic environments--including both inter-regional adjustments and adaptation to relatively small increments in local elevations (Hartwell, 1982). Documentary sources for the period and affected regions are quite rich. 3. Between 1050 and 1250, there was a second transformation of agriculture and land use in the Lower Yangzi core characterized by widespread empolderization and conversion of wetlands into cultivated acreage (Steurmer, 1980). In addition, extensive areas of upland paddy were converted to cotton fields during this period. The extant data available for use in reconstructing and mapping these transformations are enormous and includes micro-data derived from stele recording school land information that will allow the location of these changes field by field and farmer by farmer in some areas. 4. Vast areas of Northwest China, Manchuria and Sichuan were transformed into wasteland frontiers between 1250 and 1400--partly owing to famine and dislocations associated with Mongol campaigns in China and partially to widespread epidemic in these regions as well as in Southeast Coast counties. Some of the most densely populated regions of China in 1150 were converted into sparsely populated frontiers by 1400. Total population fell from about 150 million in 1150 to not much more than 80 million in 1381. Sichuan's census records show a drop from more than 16,500,000 people in 1150 to less than 1,500,000 in 1381. By 1580, recovery resulted in a regional distribution and total population that was almost an exact replica of what it had been in 1150 (Hartwell, 1987). 5. The possible expansion of settlement into Lingnan and continued economic growth in the Middle and Lower Yangzi physiographic regions were abruptly interrupted by the 12th and 13th century demographic crises, and lower costs in recovering abandoned lands as compared to reclamation of new lands delayed Lingnan development for several centuries. Sometime in the first half of the sixteenth century, economic activity in many parts of China, but especially in the lower and middle Yangzi regions and in Lingnan, accelerated at a rate faster than the growth in population. By all counts, during the eighteenth century, per capita productivity increased at the same time that population grew from 150 to 300 million souls. Developments in Lingnan between 1550 and 1800 provide particularly rich data for testing models of changing production functions during periods of rapid agrarian change. Economic growth and widespread impoundment, in some areas accompanied by 10-fold increments in population, transformed Lingnan from a frontier to a densely settled multiple cropping paddy region. Between 1500 and 1800 population growth in regions characterized by two full crops of staple rice was 1 1/2 time as great as the norm (Marks, 1991; Marks, forthcoming; Hartwell, 1987). 6. High rates of growth, 1500 to 1800, also characterized some regions in which new world crops were most common, but the effects of these innovations had their greatest impact in the subsequent period. Double-cropping rice regions continued to post above average rates of population growth during the nineteenth and twentieth centuries, but areas of West and North China where new world crops had their greatest influence (Ho, 1959) also reported some of the highest rates of increase--nearly 1 1/2 times the norm. Of particular importance were those areas in which maize became a summer grain planted after the harvest of a winter wheat crop. These regions averaged growth rates of .38 % per annum between 1800 and 1950, or 1.4 times the norm. A band of territory, which recorded an average of .56 percent a year rate of population growth during this 150 year period--some regions registered nearly 1% annual increments in population--was characterized by a growing season too short to double crop wheat with kaoliang or millet, but long enough to double crop wheat with maize which has a slightly shorter growing season than either of the two traditional grains (Hartwell, 1987). The effect was a massive clearing of hitherto sub-marginal lands which could now support two crops a year. Macro-analyses of these large-scale transformations define the broad parameters for micro-analytical case studies, whereas the latter provide detailed empirical bases for testing causal hypotheses designed to explain the provenance and effects of the major trends. CITAS plans to create over the next five years a decision-support delivery system that will both facilitate these analyses as well as help examine and refine existing decision- support systems. The adaptation, modification and refinement of existing decision-support technology as the primary user interface with the CITAS database as opposed to the adoption of current or the creation of new models of, e.g. land-use/cover change, was a conscious choice. Decision-support systems are theory neutral. Each researcher can use a transparent and integrated set of GIS, statistical and simulation tools as a means to test out alternative configurations of data under different assumptions about desired and undesired results. A retroactive decision-support system, based on the most advanced computer technology, and utilizing a bank of data on environmental change over fifteen centuries should revolutionize research on the human dimensions of global environmental change as well as facilitate the development of more rigorous social science longitudinal theoretical models. Over the next five years, CITAS intends to build a database and user interface that will allow any investigator to select any period of time and any set of variables included in the CITAS database in order to test any model of change. A retroactive decision-support system designed to examine past decisions will yield vital information about the determinants of "wrong" versus "right" choices under controlled circumstances. Was the decision to encourage adoption of new rice varieties, thereby dramatically changing land-use/cover in many areas of South China, a "correct" one? Was it "correct" in terms of the values and objectives of the policy-makers--e.g. maximize state revenues by enhancing the tax base? If not, what determined an incorrect decision--lack of knowledge, coalitional advantage, cultural norms, etc.? Was the decision "right" or "wrong" in terms of some other alternative assumptions concerning social goals that did not play a role in the contemporary decision-making process--e.g. preservation of bio-diversity? In other words, what current models would predict specific undesigned effects of these choices? Did the empirical results, in fact, coincide with the predictions of these models? Standard evaluation models can be tested against past experience in retroactive environmental impact studies. For example, the database will include information on a region's biota, hydrology, existing land-use and land cover, economic productivity, transportation grid, climate and topography. Using a decision- support system, a researcher can use GIS, computer simulations, and statistical models to predict how a decision to introduce a new agricultural practice, e.g. impoundment, should have affected bio- diversity, regional economic development, albedo, and climate at diverse points in time, in different regions, and under different socio-cultural configurations and then conduct crucial experiments by fitting the predicted results with the empirical data derived from the documented effects of such decisions, not only over years or decades, but over centuries. And each investigator can select, without any built-in assumptions, from an extensive body of data in order to test alternative hypotheses and models. The range of possible research will only be limited by scholarly expertise, imagination and levels of funding. *Robert M. Hartwell is Professor of Chinese, Department of Asian and Middle Eastern Studies, University of Pennsylvania and a member of the Policy and Planning Board of CITAS. China in Time and Space (CITAS) was founded in order to inventory, acquire, develop and disseminate electronic databases on contemporary and historical China. CITAS began as a partnership between China specialists, initially sponsored by the ACLS/SSRC Joint Committee on Chinese Studies (JCCS), and the Consortium for International Earth Science Information Network (CIESIN). During 1993-1994, GIS development projects included 1) acquisition and creation of digitized base maps of contemporary China and 2) reconstruction and documentation of historical changes in administrative boundaries as the initial basis for the development of a cross-temporal GIS. The initial GIS database will include a 1990 county boundary map of China matched to the DCW. CITAS will also have coded and documented year-by-year co-locations of changes in administrative units, county-level and higher, from 741 to 1949 as well as year-by-year changes for most years since 1949. The translator year(s) for these co-locations is 1990 so as to key the historical codes to the 1990 county boundary map. By the end of calendar year, 1994, CITAS will have extensive census and economic data for the years 1982, 1985, 1990 as well as economic data for a number of other years since 1949. The CITAS database will also include a geo-referenced collective biographical/genealogical set of 21,500 biographies containing historical demographic, migration and political variables for the period before 1644 as well as geo-coded census data for the period 741-1643; and geo-coded land use data for Lingnan during the Qing dynasty (1644-1910). CITED PUBLICATIONS Hartwell(1982) Robert M. Hartwell, "Demographic, Political and Social Transformations of China, 750-1550," Harvard Journal of Asiatic Studies (December, 1982), pp. 365-442. Hartwell(1987) Robert M. Hartwell, "Societal Organization and Demographic change: Catastrophe, Agrarian technology, and Interregional population trends in Traditional China," 2nd Congres international de demographie historique, Paris, France, june 4-5, 1987. Ho (1956) Ping-ti Ho, "Early-Ripening Rice in Chinese History," Economic History Review (December, 1956), pp. 200-218. Ho (1959) Ping-ti Ho, Studies on the Population of China, 1368-1953 (Cambridge: Harvard University Press, 1959). Kato (1954) Kato Shigeru, Shina keizaishi kosho [Studies on Chinese Economic History] (Tokyo, 1954), II, 638-675. Marks (1991) Robert B. Marks, "Rice Prices, Food Supply, and Market Structure in Eighteenth-Century South China." Late Imperial China 12.2: 64-116. Marks (forthcoming) Robert B. Marks, Environment and Economy in Late Imperial South China, 1400-1850. Perkins (1969) Dwight H. Perkins, Agricultural Development in China, 1368-1968 (Chicago: Aldine Publishing Company, 1969). Steurmer (1980). John Steurmer, "Polder Construction and the pattern of Land Ownership in the T'ai-hu Basin During the Southern Sung Dynasty," Diss. Univ. of Pennsylvania, 1980. Zhang, et. al (1993) Zhang Peiyuan et. al., "Climate Change and It's Impact on Capital Shift During the Last 2000 Year in China," Paper presented at Conference on the History of the Environment in China, Hong Kong, December, 1993. Zhang and Zuo (1990). Zhang Peiyuan and Zuo Dakang, "The Huang- Huai-Hai Plain" in B.L. Turner II et. al., The Earth as Transformed by Human Action (Cambridge: University Press, 1990), pp. 473-477.