Real estate agents know that the value of a property can increase or decrease significantly due to three important factors: location, location, location. The UK’s Channel 4 broadcasts a real estate programme using exactly this catchy phrase and each episode is fully focused on finding the perfect home in the perfect location. The problem with a house, of course, is that it cannot be moved; it is fixed at a specific location. So, the perfect home in the wrong place suddenly becomes a lot less valuable, and the wrong house in the perfect location leaves the owner with a lot of extra work (and additional investments in terms of time and money) to turn the house into a real home.

Interestingly, the same can be said for water quality sensors at a water utility. Sensor technologies can provide a wealth of water quality information, but if the sensor is installed in the wrong location, the data suddenly become a lot less valuable, as it cannot be translated into process performance. By contrast, the wrong sensor in the right location leaves the utility with a lot of extra work (and additional investments in terms of time and money) to turn the collected data into actionable information. It is, however, very surprising how little effort is made to select the right sensor technology in relation to the monitoring purpose and determine where and how often the sensor should collect data points. A business case for online monitoring very often does not go beyond determining the parameter(s) to be monitored online in order to fill the gap between sampling events. Many utilities decide to switch to online monitoring in order to increase monitoring frequencies for the purpose of improved process control. As this is often not economically feasible by means of sampling and laboratory analysis, an online sensor is suggested to solve the issue. So far, so good.

But when the ‘how’ question needs to be answered, many utilities automatically turn to the sensors they already know (even if their performance is not optimal) or put out tender procedures without specifying exactly what their needs and requirements are. Especially when online sensors are part of larger (infrastructural) projects, the performance of the sensor technologies in relation to required specifications rarely influences the buying decision. One argument often heard is: “We always buy from the same supplier, because this simplifies maintenance contracts and spare supplies.” Of course, this is true, but it also sounds as if convenience is more important than water quality. If insight into water quality is the main purpose of installing sensor technologies, then we should expect the selection process to reflect this. The number of commercially available sensor technologies for water quality monitoring is increasing rapidly by more than 100 new sensors per year, and this leads to more and more niche-instruments becoming available for a variety of specific purposes. It is therefore worthwhile spending some time to explore this market, as there may be instruments available which fit your needs exactly, but may not have come to your attention yet.

The ‘where’ question often receives even less attention. The location of a sensor is mostly determined by practical issues. Of course, it is easier to install a sensor at a surface water pumping station than in the middle of a lake, as there is grid power available and the station is visited by (maintenance) staff on a regular basis. But is that pumping station really the best location from a water quality perspective?
Again, convenience seems to be more important than water quality here. We need to realize that (management) decisions will have to be taken on the basis of the water quality data collected from this point, and representativeness of a monitoring location is rarely reported and thus rarely assessed. This can have far-reaching (and expensive) implications.

Luckily, there is a major difference between real estate and sensors: sensors are movable objects, so if it is in the wrong location, move it! This adds, however, an additional question: should the sensor be installed at a fixed point or be used as a mobile device? The answer is neither simple, nor straightforward. Some parameters are best measured at a fixed location, so that trends over time can be observed. Other parameters provide much more insight if multiple measurements are obtained over a larger area, so that geographical variability can be assessed.
Until not so long ago, we had to depend on locations where grid power and communication options were available in order to install sensors and retrieve data. However, modern off-grid power supplies and communication protocols have given us the freedom to place sensor technologies where they are most needed. Provided the right sensor is chosen for the job, the reliability and thus the value of the collected data can be increased significantly.

The key to putting the right sensor at the right location is to follow a coherent stepwise approach towards selecting and installing online monitoring equipment, starting from a carefully described objective and taking into account all relevant conditions and circumstances affecting data quality. This requires time and effort in the preparatory phase. In my experience, however, it repays itself tenfold or more when operating the sensors and converting its data into meaningful information.