1 Kenya Meteorological Department P O Box 30259-00100 Nairobi, Kenya

OBSERVED SIMILARITY IN SATELLITE DERIVED TOTAL COLUMN OZONE DATA AND GROUND BASED DOBSON AND OZONESONDE OVER NAIROBIKENYA

C. C. Okuku1 N.J Muthama2 F J Opija2

2 University of Nairobi, Department of MeteorologyP O Box 30259-00100Nairobi

This study sought comparison of satellite total column Ozone data from 1985 to 2003 against ground based Dobson spectrophotometer and Ozonesonde data. The area of study was Nairobi and the main objective was to ascertain the similarity of the three total column ozone data sets. The inter-comparison was carried out by paring two data sets of corresponding Julian day. The root mean square (RMS) error, bias and percentage difference were used in order to achieve the objectives of the study. The RMS error for Dobson/satellite was between 0% and 26%, percentage difference with Dobson as reference ranged from -6% to 20%. Satellite/Ozonesonde and Ozonesonde/Dobson data sets showed variable RMS error and percentage difference. Bias was positive on average ranging between 2du to 12du in all the three data sets.Dobson/satellite results showed small differences of RMS error, bias and percentage difference hence were comparable. The ozonesonde data had some discrepancies partly because of vertical dynamics of Ozone and that of the balloon carrying Ozone sensors

INTRODUCTION

Public concerns about human impacts on the earth’s atmosphere have been growing because of Ozone layer decline, climate change and air pollution increase. It is important to observe Ozone’s vertical distribution on a global scale. From the ground, the Ozone layer can be monitored very accurately with a high vertical resolution using balloon sondes, Dobson spectrophotometer and laser radars. However, extending these high-quality data to a global scale requires space-borne instruments e.g. satellites.

Therefore there is need for comparison of these data sets, which this study attempts to do.

OBJECTIVES

The aim of this study is to assess the similarity of the observed satellite Ozone data and ground based i.e. Dobson spectrophotometer and Ozonesonde data over Nairobi, Kenya.The specific objectives are,

To compare the Ozone data derived from satellite overpass and ground based Dobson spectrophotometer and Ozonesonde data.
To determine which data set is more representative over Nairobi.

JUSTIFICATION OF THE STUDY

Models of the world’s climate require, among others, the knowledge of the tropical atmosphere of the East Africa, because of its position with respect to the high pollution sources represented by countries facing the Indian Ocean. These models need data from space-borne and balloon-borne sensors since repetitive, accurate and global physical measurements are necessary to quantify atmospheric processes and to improve the understanding of global climate dynamics.

The tropical variation of atmospheric ozone is of crucial importance for the radiative balance of the Earth-Atmosphere system and help in detecting long-term trends of major importance.

Despite the small amount of ozone in the atmosphere, it has several vital roles. These include:

Ozone helps in maintaining the stratospheric temperature distribution. This warmth is due to absorption of solar ultraviolet (UV) radiation by the ozone. A reduction of ozone concentration in the stratosphere has the effect of lowering the temperature in this region (IPCC, 1994).

LITERATURE REVIEW

In 2004 Lahnemann during his under graduate project at the university of Nairobi, Department of physics compared the Differential Optical Absorption Spectroscopy DOAS, record with the data from TOMS, Global Ozone monitoring experiment, GOME and the Scanning Imaging Absorption Spectrometer for Atmospheric Chartography, SCIAMACHY, as well as the Ozonesondes of Nairobi and found a very good agreement in trend for all instruments, the root-mean-square values of the deviations from DOAS were all below 3.5% and the mean deviations were even less (< 1.5%). Correlation between the different data sets was also very good for most of the time observed, with a deviation < 5%.

In 2002 Fioletov et aal carried out validation of ground based and satellite measurents, 1964-2000 and found that the data over tropical and equatorial region and Southern hemisphere agree within 2-3%.

It is very desirable to keep up the Nairobi DOAS station for Dobson spectrometers have yielded a very valuable record of total column ozone in many places around the world, but the clear advantage of DOAS is the different absorbers it can observe. As tropical data is still scarce and many more stations are operated in mid and high latitudes a special focus should be laid on maintaining the observations in Nairobi. The launch of SHADOZ ozonesondes nearby provides another valuable addition, ‘Validation of SCIAMACHY TOSOMI ozone columns withGround based data (draft version) Ellen Brinksma, KNMI, De Bilt, the Netherlands,2006’

Stratospheric ozone depletion and its subsequent impact on the biosphere has gained considerable attention within the international scientific community in the recent past. Scientific evidence has shown that human-made chemicals are responsible for the observed depletion of the ozone layer (IPCC, 1994). These chemicals contain various combinations of the chemicals elements chlorine, fluorine, bromine, carbon and hydrogen.

DATA AND METHODOLOGY

The data that was used includes measured total column Ozone values using Ozonesonde at Kenya meteorological department headquarter, KMD, (Dagoretti Corner) between 1998 and 2003, Dobson Spectrophotometer data from 1985 to 2001 and satellite overpass data from 1985 to 2003. Ozonesonde data is only measured on Wednesdays

METHODOLOGY

Comparison is the act of examining resemblances. In this study comparison is used to mean the relation based on similarities and differences.

Firstly, the three sets of ozone data were collected i.e. Ozonesonde, Dobson Spectrophotometer and Satellite. Corresponding dates isolated for all the years using Julian days.

The second step now separates the error of reference data from the difference between the ground based observations and the satellite. This is achieved in three levels, as follows;

Estimation of bias

Bias B

O is the observed E is the reference .

Apositive bias indicates that the reference exceeds the observed value on the average. Systematic errors are unearthed by the results of the bias if for a given station the bias values bear the same sign. In this study ground based data was treated as the observed and satellite as reference data

Error separation method (ESM)

It follows the Ciach and Krajweski (1999). It is based on error variance and is described by Root mean square RMS error.In mathematics, root mean square also known as the quadratic mean, is a statistical measure of the magnetude of a varying quantity. It is especially useful when variates are positive and negative, e.g. waves.

Percentage difference

In this study percentage difference was obtained using the following formula

Observed-Reference)/Reference * 100

RESULTS AND DISCUSSION

The first pair of data set to be analyzed was Dobson spectrophotometer and Nimbus-7 satellite from 1989 to 1993 (and named category 1) when it stopped operations and used Toms/EP satellite data period 1996 -2003 ( category 2). This division of analysis into two was necessitated by the fact that there was no satellite data from Nimbus-7 the years 1994-1995 which coincided with the long of the second generation satellites. Also this was to allow see any change in the satellite data. See table 1 below for the results

Table 1 Results of Dobson spectrophotometer and Nimbus-7 satellite

RMS error / 0%-26%
BIAS / 0 –13 Du
%Difference Dobson / -6%-21%
%difference satellite / -17.2%-10%

Figure 1 percentage difference of Dobson spectrophotometer and nimbus-7 satellite (category 1)

Figure 1 percentage difference of Dobson spectrophotometer and TOMS-EP satellite (category 2)

Figure 2 percentage difference of Dobson spectrophotometer and Ozonesonde

Figure 3 percentage difference ofOzonesonde and satellite

Figure 3 Bias ofOzonesonde and satellite

Conclusion

In this study the three Ozone data sets i.e. Dobson spectrophotometer/Satellite, Ozonesonde/satellite and Dobson spectrophotometer/Ozonesonde, were found to have big data Variability though comparable. However, there was some good agreement in some days and months within the acceptable margins of RMS error and Percentage difference of between 2% and 8%, Muthama, 2004. This suggested similarity especially Dobson spectrophotometer andSatellite.

Recommendation and suggestions

There is need for daily reading of Dobson spectrophotometer data and the weekly Ozonesonde flights to eliminate missing data for future research.
As part of capacity building KMD should train more staff for competence in understanding and handling of Ozone data. It is unfortunate that both Ozonesonde data for years 2004 to 2006 and Dobson spectrophotometer 2005 to 2006 could not be traced. There is need for recalibration and validation of the three instruments to minimize separation error. More time is also required for Ozonesonde readings for future comparison and validation.
When using Ozonesonde data it is recommended that you filter out the extreme values due to residual Ozone interpolation.

Acknowledgement

This work was carried out with support of Kenya Meteorological Department, University of Nairobi, Kenya and Meteo-Swiss, Switzerland. The authors are also grateful to workmates for providing necessary support and advice.

Reference

Atmosphere, Weather, and Climate. Barry, Roger G. and Chorley, Richard J. Routledge, New York, 1992.
Atmospheric Chemistry and Physics of Air Pollution. Seinfeld, John H., John Wiley and Sons, Inc., New York, 1986.
Introduction to Theoretical Meteorology. Hess, Seymour L., Krieger Publishing Company, Malabar, Florida, 1979.
IPCC (2007). Intergovernmental Panel on Climate Change, fourth assessment report.
Meteorology Today. Ahrens, C. Donald. West Publishing Company, St. Paul, Minnesota, 1991 and 1994.
Muthama 1988. Master’s thesis
Ozone depletion scientific assessment report. (2006)