Is it feasible to estimate radiosonde biases from interlaced measurements?

Stefanie Kremser,Jordis S Tradowsky,Greg E Bodeker,Henning W Rust

doi:10.5194/amt-11-3021-2018

Abstract

Abstract. Upper-air measurements of essential climate variables (ECVs), such as temperature, are crucial for climate monitoring and climate change detection. Because of the internal variability of the climate system, many decades of measurements are typically required to robustly detect any trend in the climate data record. It is imperative for the records to be temporally homogeneous over many decades to confidently estimate any trend. Historically, records of upper-air measurements were primarily made for short-term weather forecasts and as such are seldom suitable for studying long-term climate change as they lack the required continuity and homogeneity. Recognizing this, the Global Climate Observing System (GCOS) Reference Upper-Air Network (GRUAN) has been established to provide reference-quality measurements of climate variables, such as temperature, pressure, and humidity, together with well-characterized and traceable estimates of the measurement uncertainty. To ensure that GRUAN data products are suitable to detect climate change, a scientifically robust instrument replacement strategy must always be adopted whenever there is a change in instrumentation. By fully characterizing any systematic differences between the old and new measurement system a temporally homogeneous data series can be created. One strategy is to operate both the old and new instruments in tandem for some overlap period to characterize any inter-instrument biases. However, this strategy can be prohibitively expensive at measurement sites operated by national weather services or research institutes. An alternative strategy that has been proposed is to alternate between the old and new instruments, so-called interlacing, and then statistically derive the systematic biases between the two instruments. Here we investigate the feasibility of such an approach specifically for radiosondes, i.e. flying the old and new instruments on alternating days. Synthetic data sets are used to explore the applicability of this statistical approach to radiosonde change management.

Highlights

Radiosondes are indispensable for monitoring the upper air as they provide high vertical resolution in situ observations of temperature, pressure, and water vapour between the surface and the upper troposphere–lower stratosphere
We investigate the feasibility of quantifying the difference in biases of two instrument types by alternating between the two different instruments and applying a statistical model to infer any systematic biases between the two instruments
Performing dual radiosonde flights with both instrument types is costly, and we investigated the feasibility of using interlaced flights to obtain an estimate of the difference in the bias

Summary

Introduction

Radiosondes are indispensable for monitoring the upper air as they provide high vertical resolution in situ observations of temperature, pressure, and water vapour between the surface and the upper troposphere–lower stratosphere. Determining long-term temperature trends from radiosonde measurements is challenging because changes in instrumentation can, among other things, introduce discontinuities in the measurement time series (see Fig. 1). Since radiosonde measurements are primarily made to provide the data needed to constrain weather forecasts and not to detect long-term changes in climate, little attention has been paid to ensuring the long-term homogeneity of the measurement record when changing from one instrument to another. Radiosonde data records typically fall short of the standard required to reliably detect changes in climate. Another cause of inhomogeneities in the record is undocumented changes in data processing (Thorne et al, 2011). While much effort has been spent attempting to remove discontinuities in radiosonde data records (e.g. Sherwood et al, 2005; Randel and Wu, 2006; Haimberger et al, 2012), lack of confidence in the long-term homogeneity erodes confidence in derived trends. Seidel and Free (2006) used upper-air temperatures from the NCEP-NCAR reanalysis (Saha et al, 2010) to in-

Methods

Results

Conclusion