Premium
How Temperature Sensor Change Affects Warming Trends and Modeling: An Evaluation Across the State of Colorado
Author(s) -
Ma C.,
Fassnacht S.R.,
Kampf S.K.
Publication year - 2019
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1029/2019wr025921
Subject(s) - climate change , environmental science , snow , temperature record , climatology , homogenization (climate) , elevation (ballistics) , weather station , calibration , global warming , air temperature , mean radiant temperature , atmospheric temperature , meteorology , statistics , geography , mathematics , geology , biodiversity , ecology , oceanography , geometry , biology
In the western United States, a temperature sensor change at the snow telemetry stations is responsible for erroneously greater temperature warming trends in high‐elevation mountain areas than lower elevation locations. This study examined how the temperature sensor changes influenced these trends across Colorado and evaluated two homogenization methods that adjust these data biases. Perspectives differ on whether the presensor (before the changing of the temperature sensors) or post‐sensor (after) change data are correct, so temperature records from both the pre‐ and post‐sensor change period (~1990s to mid‐2000s) were individually adjusted at 68 longer‐term stations. Trends were analyzed with and without the adjustments using the Mann‐Kendall significance test and Theil‐Sen's rate of change. Initially the post‐sensor change data were used to calibrate a temperature index snow water equivalent model that was evaluated with the original and adjusted temperature data sets. Mean temperature warming trends for the original data set averaged 0.95° per decade and reduced to less than 0.5° per decade for the adjusted data sets. Results from the temperature index model showed that snow water equivalent was simulated better with the homogenized temperatures from both methods relative to the original temperatures, with 44–69% of the stations within “good” and “very good” performance categories. This modeling was repeated using calibration of the presensor change data yielding very similar results. These findings show that accurate reconstruction of the historical temperature records is challenging but temperature adjustment methods can create more reliable temperature records for climate change analysis and hydroclimatic modeling.