Value Added Abstracts
Naiyana Sahavechaphan, Jukkrap
Abstract
Weather observations have played an essential role in several domains such as agricultural and environmental sciences. In modern weather stations, observations are automatically sensed and transmitted via a wireless link to a centralized database in a timely-fashion. Each individual station, in particular, sits on its own surrounding area and consists of (i) meteorological sensors to measure environmental observations (i.e. temperature, rain and humidity); (ii) mainboard to control the work process; (iii) energy source (i.e. solar cell and battery) to drive the work process; and (iv) communication network to transmit observations from field to server. Here, it is likely that sensors can malfunction, battery degrades, communication is poor, rain gauge is blocked and etc. Consequently, observations could be anomalous, delayed or missing. It should be noted that the more accurate, complete and in-time observations, the more effective and reliable are their applications. It is thus vital to always maintain weather stations such that the accurate, complete and in-time observations are promising. Here, we term this promising as weather station health. We believe that the health of weather stations should be real-time evaluated and then reported to the corresponding officials for timely maintenance plan and action. In this work, we thus propose TanPibut HEALTH which is a software system that holistically evaluates the health of weather stations. Specifically, it evaluates a set of observations starting from the current moment back to a specified period of time based on the qualified, missing and delayed observations. The evaluation result is given with the numerical value ranging from 0.0 to 1.0 for the worst to the best station health.