|
In the 1980s when Dr. Lansing did his initial research, he worked with other scientists to create a computer model of Balinese agriculture. To understand what this computer model demonstrates, read on for excerpts from the original documentation (with our emphasis). Interact with a web enabled version of the scientist's Computer Model of Balinese Agriculture. Excerpts from "The Formulations of the Bali Model" prepared by James N. Kremer (Draft Version 1.1: January 1989)
Background
Dr. Steve Lansing had been working on the role of the Balinese Water Temple system in rice irrigation since 1983. In 1986, a USC grant provided travel funds allowing Dr. James Kremer to accompany Lansing to Bali to assess first hand the prospect for developing a computer model of the system. Based on the hydrological and agricultural data they obtained, plus the anthropological evidence from Lansing's earlier work, an NSF proposal was submitted and funded jointly through the anthropology program.
The original goals of the NSF research were to develop a computer simulation model of hydrology, rice growth, and pest population dynamics. Using this model of the watershed, various management scenarios could be explored. The hypothesis to be tested was whether an optimum scale of regional coordination existed to strike a balance between two conflicting goals: efficient use of water and control of insect and insect-borne virus pests.
Local cooperative groups of farmers called subaks always plant and harvest according to the same schedule. But the degree of coordination among hundreds of subaks up and down the river basin is the key. During the dry season, if all subaks stagger their planting schedule, water demand is spread out, making efficient use of the water supply. But pests developing in one field can readily move at harvest time to an adjacent field if that field is on a different schedule. Project reports from the Bali Irrigation Project funded by the Asian Development Bank documented the increased crop loss due to pest damage that resulted when tulak sumur ("every man for himself") was practiced. At the other extreme, were all subaks to plant in synchrony, pest populations would be reduced across the whole landscape by the fallow period between the multiple crops. In this case, however, water supplies would be inadequate and this resource would be inefficiently used.
The simulation model was developed specifically to determine if there was an optimal spatial scale of coordination to balance the water use and pest control. Results to date suggest that such an intermediate scale is indeed optimum and that it corresponds to the scale of coordination that results from the structure imposed on the system by the network of Water Temples, and administered by the Water Temple Priests.
Geography
The Gianyar region of Bali includes the watersheds of the Oos and Petanu Rivers. We divided this watershed into 12 sub-sections specifying the catchment basins for each of the dams for which hydrological data were available. The area, elevation, and proper spatial connectivity of the basins, rivers, dams are known from Balinese Irrigation Project (BIP) reports, confirmed and supplemented with maps and data we obtained from Balinese Departments of Hydrology and Public Works.
For the 172 subaks of the region, we specify their name, area, the basin in which they reside, the dam from which they receive irrigation water, and the dam to which any excess is returned. We also specify the real spatial mosaic connecting these subaks. Thus, for each one we specify the neighboring subak on all four sides or if it is bounded by another kind of boundary, like a river, road, or city. Cropping patterns are eventually specified for each of these subaks, and each one potentially functions separately with respect to cropping schedule, water demand, crop growth, and the immigration, emigration and growth of the pest population.
Crop Model
By choosing among the various management scenarios, the user specifies crop planting patterns to be followed by each of the subaks. The growth dynamics for the crops are presently formulated in a very simplistic way designed to focus directly on the rice harvest and its reduction by water stress and pest damage. Detailed physiological data that might be used to formulate a mechanistic rice model are not available for the traditional varieties, and in any case, such a detailed model is inappropriate for our purposes.
The model presently provides 5 crop options including a fallow period, vegetables, and 3 rice varieties. Padi Del is the traditional variety planted as the first crop. It is planted in the rainy season and matures 5-6 months later about a month into the dry season. The traditional second rice crop is Padi Chi Chi that matures in 4 months. The High Yield Variety is assumed to mature in 3 months. For each crop, data specify the time from planting to harvest, the normal maximum yield, and the potential effect on yield of severe pest damage and of withholding fertilizer.
Rice Stage |
