HYDROLOGIC DATA, GUANXI, AND GROUND-WATER MODELING IN NORTH CHINA: THE PRICE OF INFORMATION

Every five days since the mid 1970s, workers have measured and recorded the depth to water in 14,000 monitoring wells throughout North China. During the period of record, water levels have dropped precipitously and water-use patterns have changed. Together, the water-use changes and their recorded impacts make ground-water modeling an ideal tool for examining the underlying causes of the water-level declines and testing ideas for reversing them.

However, the data are extremely difficult to access. Economic reforms in the early 1990s significantly reduced government funding for provincial water conservation bureaus, which maintain water-level databases. Lacking other income sources, the bureaus sell data to support their activities. At 15 yuan (about US$2) per measurement, existing water-level data to calibrate a 30-year model of a small (400-km²) watershed are priced at more than US$1 million. Competing missions between government agencies further complicate data acquisition. For example, an agricultural bureau may be unwilling to release flow and chemical data for an irrigation system if its water quality exceeds standards set and enforced by an environmental bureau. Even within agencies, hierarchical culture and social norms discourage data sharing. As a result, guanxi, an informal but deeply ingrained system of personal give-and-take relationships, has become the only viable means to obtain data.

Hydrologic modelers who lack guanxi compensate by making assumptions about the systems they model. In one case, American and Chinese counterparts took different approaches to compensate for missing data, and consequently modeled two distinct hydrologic interpretations of the same system. Not surprisingly, their management recommendations also diverge. Data deficiencies forced model truncation at political rather than hydraulic boundaries, so boundary conditions for each timestep had to be input explicitly. As a result, neither model is suitable for predicting future conditions. The high cost of data, in terms of either money or guanxi, leads to shortcuts that compromise the utility of hydrologic models, just when models are urgently needed to evaluate and guide water-management changes in North China.