Verification of LTPP Virtual Weather Stations Phase I Report: Accuracy and Reliability of Virtual Weather Stations

Introduction

Several climatic measurements are taken during the Long-Term Pavement Performance (LTPP) program. Onsite weather data are collected during the Seasonal Monitoring Program (SMP) for 63 General Pavement Studies (GPS) and Specific Pavement Studies (SPS) sections. Climatic data are also collected for 35 SPS–1, –2, and –8 sections using onsite automated weather stations (AWS). In the Information Management System (IMS) database, the climatic conditions for 880 SPS and GPS sites are estimated using data from as many as 5 nearby national weather stations. These estimates are referred to as virtual weather stations (VWS).

The study of onsite and estimated climatic data in the LTPP database can help determine the reliability and accuracy of climatic data and identify any possible discrepancies in the data. It will also help determine the quality of VWS estimates and the accuracy with which the site’s weather conditions can be estimated.

BACKGROUND

The climatic database for the LTPP program was originally developed by the Strategic Highway Research Program (SHRP) in 1992(1) and was later revised and expanded by the LTPP program in 1997.(2) The climatic data in the LTPP database were derived from the U.S. National Climatic Data Center (NCDC) and the Canadian Climatic Center (CCC) databases. For each GPS test section and SPS project site, up to five nearby weather stations were selected and used to estimate site-specific climatic conditions. These estimates are referred to as VWS. The VWS climatic data are estimated by averaging data from up to five nearby weather stations (see figure 1) using a 1/R2 weighting scheme as shown in figure 2.

Figure 1. Diagram. Schema of VWS concept.

Figure 2. Equation. Estimating climatic data.

where:

Vm = Calculated data element for day m for the VWS.
k = Number of weather stations associated with the project site (up to five).
Vmi = Value of a data element on day m for weather station i.

Ri = Distance of weather station i from the pavement project site.

The monthly and yearly estimates of several climatic parameters for every LTPP section are developed and stored in several tables in the climatic module (CLM) of the LTPP database. Between 13 and 57 years of data through 1996 are available for the LTPP test sections.

OBJECTIVES

The objectives of the study are to:

• Characterize the variation between the estimated virtual weather data and onsite measurements to determine the accuracy of virtual weather data and compare it with the accuracy of within-site measurements and yearly variation.
• Investigate factors affecting the VWS estimates and propose changes to the existing algorithm that can improve the accuracy of the estimates.
• Report any anomalies found in the measured and estimated weather data during the conduct of the project.

TASKS

The project was conducted in four tasks:

Task 1:

Obtain and summarize climatic data.

Task 2:

Compare estimated and onsite climatic data.
Compare daily temperatures.
Compare monthly temperatures.
Compare monthly precipitation.
Compare monthly freezing index.
Compare all other parameters.

Task 3:

Determine the accuracy of VWS estimates and verify the VWS algorithm.
Identify suspect data (outliers).
Determine the accuracy of VWS data.
Determine the factors affecting VWS estimates.
Recommend modifications to the VWS concept.

Task 4:

Prepare Phase I report.