Meteorological and hydrological sensors deployed over several hundred kilometers of geographical area comprise an environmental sensor network. Large amounts of data need to be processed in minimal time and transmitted over the available low speed and low bandwidth links. This paper describes algorithms for optimal data collection and data fusion. An inductive model using exponential back-off policy is used to collect optimal amount of data. The data measurements for temperature, pH and specific conductance collected for a year from the sensors deployed at Lake Lewisville are used to test the inductive model. Energy savings of 90% are achieved even with 1% of degree of tolerance. The problem of data fusion is addressed by the introduction of a novel concept of a super-sensor, based on self-organization and collaboration among sensors. A histogram application is described that uses recursive doubling for global collaboration between sensors. The performance of the networked super-sensor in comparison to a centralized polling approach is analyzed for optimality on two different geographical areas. ©2004 IEEE.