UK Radiosonde testing 2002/2003 and improvements to the Camborne test facility

1 00/XXXX © Crown copyright

UK Radiosonde testing 2002/2003

and improvements to the Camborne test facility

Richard Smout, John Nash

•CIMO ET on Upper Air System Comparisons – Geneva 17-20 March 2003

•Agenda item 3.2


Summary of testing

2001 military radiotheodolite evaluation 2002 RS90 humidity (February) 2002 RS92 high power (October) 2003 RS92 low power (May) 2003 Meteolabor Argus 37 (July) 2003 RS92 pre-acceptance test

(December)


Until 2003 a Cossor WF Mk4 wind finding radar (S band) was used as a height and wind measurement reference.

During 2003 the radar was removed allowing the room to be refurbished for use as a dedicated radiosonde test facility with network connections to Exeter.

Three GPS radiosonde systems are permanently installed to provide the replacement working reference for height and winds.

Sippican W9000 GPS Graw GK90C GPS Vaisala DigiCORA III A Meteolabor Argus 37 is on order, intended for portable

use at international comparisons that require Snow White chilled mirror radiosondes to be flown.

Introduction to Camborne test facility


In May 2001 the Met Office tested the performance of a military mobile radio theodolite system.

The winds were checked against measurements from a Graw GPS system.

The Graw radiosonde was chosen as the GPS system easiest to operate in the field. Winds had been shown of good quality at earlier tests at Camborne.

Recent tests in the UK


The military radio theodolite system was about 10 years old. It had been serviced by the military but not the manufacturer. The particular radio theodolite tested was found to be lacking in sensitivity and did not track correctly at long ranges. Problems with operational practices were also identified.

As a result a new system will probably be purchased where a better service link to the manufacturer will be required.


BMETS wind finding results

At elevations lower than about 12º, the radiotheodolite wind speed errors increased rapidly and wind direction was also corrupted.



Re-alignment improved performance

Graw

RS80

BMETS


Radiosonde tests in 2002 Since 2000 the Met Office have been collaborating with

Vaisala on improving quality of relative humidity measurements from the RS90/92 radiosonde under funding provided from the Met Office next generation radiosonde project.

Areas examined were

1. Improved protection from contamination in cloud for relative humidity measurements.

2. Improved calibration of humidity sensor

3. Wind finding performance with code correlating GPS system.

4. Performance of radiosonde links with new RS92 transmitters.


RS90 humidity evaluation

When a tube was placed over the humidity sensors to eliminate contamination caused by water droplets on the sensors while passing through thin moist level cloud it was found that heat from the upper pulse heating sensor was ducted over the lower sensor causing it to read too low.


Radiosonde tests in 2002

Vaisala RS90 humidity test February 02 The test re-evaluated the effect of placing

a cap or tube over the RH sensors. To prevent heating of the upper sensor

influencing the lower sensor only the lower sensor was heated.

The uncapped RS90 responded faster than the capped variant in the upper troposphere.


Comparison of Humidity Response in Medium to High

Cloud

CappedRs90

UncappedRS90

RS80



Throughout the trial the uncapped RS90 with both sensors pulse heated responded quicker than the capped RS90 on emerging from high cloud

Response in low cloud was very similar for the two systems


Precipitation contamination

During flights through precipitation the uncapped RS90’s pulse heating , at a faster rate than for a standard RS90, was sufficient to avoid contamination.

The capped RS90 and RS80 both suffered from contamination.


RS80

CappedRS90

UncappedRS90


Conclusions Met Office forecasters require cloud tops

accurate to better than 100 meters for low cloud.Both capped and uncapped RS90s fulfilled the requirements for detecting cloud tops (given a modified heating cycle). The uncapped RS90 was the preferred option due to its faster speed of response at upper levels.

It was found that pulse heating applied to the RS90 upper sensor affected the measurements of the lower sensor on some flights.

This problem could be avoided if the sensor boom were fixed at a suitable angle, rather than being adjusted at the discretion of the operator.


RS92 evaluation

In October 2002 a preliminary test of prototype RS92 was performed.

The sensor boom was now held at 45º eliminating the problems caused by warmed air from the upper humidity sensor heating the lower sensor.


High Power RS92 evaluation

Data transmission and availability was evaluated.

Humidity sensor and pulse heating performance was examined.

The RS92 now had a full code correlating digital GPS receiver which was also evaluated.


Humidity Sensor Performance

To overcome the problems reported in previous trials Vaisala had modified the pulse heating cycle to start when the temperature fell below 3ºc and stopped at -65ºc.

This initial test was successful, data reception was very good and some of the humidity contamination issues had been addressed.



Further evaluation of prototype RS92 occurred in2003. Vaisala designed a low power version. To test this new radiosonde required the loan of a digital

receiver and pre-amp for the RB21. The test showed satisfactory performance. A comparison

between high and low power RS92s showed that both types of radiosonde were transmitting data reliably , but it also showed some deficiencies in the various antenna used for testing at Camborne


Comparison of RS80, RS92 AGP & SGP data loss

RS80RS92 AGP

RS80RS92 SGP

PTU 5.21 % 3.69 % 3.5 % 6.54 %

WIND 6.37 % 0.43 % 1.06 % 0.16 %

RS92 AGP – low power, RS92 SGP – high power

In each test 18 dual ascents were flown


RS92 pre-acceptance test

Compared both RS92 high and low power radiosondes and RS80s on a triple rig.

Six Meteolabor Snow White chilled mirror radiosondes interfaced to Sippican MKII Microsondes were also flown to test this ground station system prior to the main acceptance test.


Testing of operational prototype RS92 GPS wind resultsThe RS92 low power measurements were taken as an arbitrary

reference, and compared against RS92 high power results and RS80 Loran (operational ) results.

RS92High power RS80

LORAN

V ms-1


Testing of operational prototype RS92 GPS wind results

U ms-1

RS92High power

RS80LORAN


Comparison of RS80, RS92 AGP & SGP data loss

In this latest test, data losses from the two types of RS92s were compared against the operational RS80 .

Both PTU and wind have been combined.

RS80 RS92 SGP

RS92 AGP

Total 1.06 0.14 0.12

Mean 0.13 0.02 0.02

SD 0.10 0.01 0.03


RS92 pre-acceptance test The next slide shows a Snow White working

well when compared against both RS92s and the RS80.

The dry atmosphere has caused Snow White to loose the thin film from the mirror surface.

The results from the RS92 testing are now considered satisfactory.

Consequently a full acceptance test of production radiosondes is scheduled for May / June this year.


Thin film lost in dry layer


Meteolabor Argus 37 Test A preliminary evaluation of the Argus 37

ground station was performed at Camborne in the summer .

The aim of this was to test the system from a deployment point of view and to evaluate the ease of use for deploying Snow White chilled mirror radiosondes, given the problems experienced at the WMO test in Brazil.


Flight 4

Snow White in good agreement with the two Vaisala radiosondes.

.


Meteolabor Argus 37 Test


Flight 7 Snow White working well when

compared against the two Vaisala radiosondes.

This a daytime flight, so that temperature differences increase in the stratosphere as deficiencies in solar radiation corrections become apparent.


Meteolabor Argus 37 Test


Meteolabor Argus 37

The Met Office is purchasing the system to support future radiosonde testing.

The system is compact enough to allow easy deployment for remote trials use.

The advanced telemetry has sufficient channels to allow comprehensive house keeping information for assisting with Snow White performance monitoring.


Future A large scale acceptance test for the Vaisala RS92

radiosonde is planned in May/June 2004, but Snow White chilled mirrors and Sippican three thermistor temperature sensing system will also be deployed to allow a better estimate of measurement accuracy

It is hoped that a collocated microwave radiometer, GPS water vapour and a cloud radar+ laser ceilometer will be used to judge the accuracy of the relative humidity measurements

UK Radiosonde testing 2002/2003 and improvements to the Camborne test facility

Documents