Abstract

Temperature data is dynamic and can be published frequently. Tools that can retrieve temperature data can be utilized on the fly to update and add-on to datasets. This type data is also collected at stations like NDAWN stations and due to the nature of that, is missing values where stations are not located. Interpolation can be used to solve this problem by creating and inferring the values in between based on the existing sample values collected. There are different methods of interpolation such as IDW, various Kriging methods, and RBF. There are differences such as exact and non-exact consideration of sample values between these different methods. They are all able to produce an interpolated surface. Overall, there is no best method, but in some studies, various types of Kriging methods have produced the “best” interpolated surfaces with considerations to the scenario variables.

Problem Statement

Temperature data is dynamic and changes over time. NDAWN provides temperature data on the web for different temporal resolutions. To examine data on a monthly frequency, 30 days of temperature data must be extracted on the fly. Temperature data like maximum and minimum daily temperatures can also be extracted on the fly to examine the high and low temperatures. The high temperature is defined in this lab as the largest value of the maximum daily temperatures recorded for the 30-day period. The low temperature is defined as the lowest value of the minimum daily temperature recorded for the 30-day period. Furthermore, NDAWN stations are dispersed unevenly. Temperatures in areas between station collection sites must be inferred through interpolation, which can be done with different methods.