This means optimizing cultivation inputs so that high yields are obtained and environmental effects are minimized. Several multinational and national initiatives aiming to improve quality of sea, lake and river water need more accurate information on effective means to decrease contaminants and nutrient discharges to waters and lower their effects, such as cyanobacteria blooms [9�C11].In order to function properly, sensor networks for water monitoring and agriculture normally require a relatively dense deployment of sensors. This leads to applications that monitor mostly local weather and soil characteristics [4]. Agricultural sensor networks have been developed for frost [3] or crop pest warning [12]. They are also an essential component in more advanced decision support systems (DSS) for crop protection [13,14].

In precision agriculture the studies have been concentrated on spatial data collection through mobile, vehicle embedded sensors or in-situ sensors deployed in the field [1]. Precision irrigation and fertilization and husbandry monitoring systems based on sensor networks have also been developed [1,15]. In water monitoring sensor networks are used for monitoring water quality and hydrology of rivers, lakes and reservoirs and for flood warning [4,5,16�C19].Although sensor networks still struggle with technical problems, such as energy-consumption, unreliability of network access and standard or software mismatches [20�C22], they have already been used for long-term monitoring under harsh outdoor conditions.

They allow monitoring remote, hazardous, dangerous or unwired areas, for instance in the monitoring and warning systems for tsunamis, volcanoes, or seismologic phenomena. The sensor webs, in turn, are an emerging technology, that is not yet in operational use outside the test beds [6].The sensor networks and sensor Cilengitide webs have a profound effect on the collection and analysis of environmental data. The data is very heterogeneous and may come from different in-situ, mobile or satellite sensors that have different temporal and spatial resolutions that may vary in accuracy and content [8]. Furthermore, the user has less control over data quality, and information needs to be extracted from a large amount of heterogeneous data. This highlights the importance of comprehensive metadata describing the sensors, data, and data quality, as well as the need for effective tools for data mining or other data gathering [4].

We present here a wireless sensor network (WSN), called SoilWeather, which aims to provide temporally and spatially accurate information, data services and (real-time) applications for water monitoring and agriculture on river basin and farm scales. We evaluate the performance of the network from the data user and network maintainer perspectives, and thus, focus on maintenance and data quality issues as well as applications.