The study presents a novel computational intelligence algorithm designed to optimise energy consumption in an environmental monitoring process: specifically, water level measurements in flooded areas. This algorithm aims to obtain a tradeoff between accuracy and power consumption. The implementation constitutes a data aggregation and fusion in itself. A harsh environment can make the direct measurement of flood levels a difficult task. This study proposes a flood level estimation, inferred through the measurement of other common environmental variables. The benefit of this algorithm is tested both with simulations and real experiments conducted in Doñana, a national park in southern Spain where flood level measurements have traditionally been done manually. the terrain. Therefore it is important to design robust software and hardware that can be adapted to any incident. † Flexibility: The network must be able to add, move or remove nodes to meet the application requirements. The network must automatically detect the changes, organising the communications in consequence. One of the most important constraints for this application concerns energy consumption. The batteries that provide power supply to these devices have, in general, a short life. To overcome this problem, this study presents an aggregation and data-fusion technique, that allows us to reduce the network's power consumption. The parameter considered in this study is the water level in flooded zones. We studied it using a simulator to determine the advantages of local processing and data fusion to reduce power consumption. The data fusion is based on a local Self-Organised Map (SOM) [5] distributed in each node of the WSN. The results obtained by simulation have been compared with the real deployment of a WSN. The experimental results in real scenarios demonstrate the performance of the system and validate the results obtained by simulation. Since the start of the installation we have collected partial information about the mounted sensors and their reliability. These results are discussed in Section 6. The rest of this study is organised as follows: Section 2 describes the application scenario of the proposed method; Section 3 summarises the state-of-the-art about aggregation and data fusion in WSN; Section 4 describes the proposed aggregation methods with outcome of this method