Giovanni Emilio Perona

A Network of Portable, Low-Cost, X-Band Radars

Radar is a unique tool to get an overview on the weather situation, given its high spatio- temporal resolution. Over 60 years, researchers have been investigating ways for obtaining the best use of radar. As a result we often find assurances on how much radar is a useful tool, and it is! After this initial statement, however, regularly comes a long list on how to increase the accuracy of radar or in what direction to move for improving it. Perhaps we should rather ask: is the resulting data good enough for our application? The answers are often more complicated than desired. At first, some people expect miracles. Then, when their wishes are disappointed, they discard radar as a tool: both attitudes are wrong; radar is a unique tool to obtain an excellent overview on what is happening: when and where it is happening. At short ranges, we may even get good quantitative data. But at longer ranges it may be impossible to obtain the desired precision, e.g. the precision needed to alert people living in small catchments in mountainous terrain. We would have to set the critical limit for an alert so low that this limit would lead to an unacceptable rate of false alarms

ROSA – The Italian Radio Occultation Mission Onboard the Indian OCEANSAT-2 Satellite

Within the next October 2008 the Indian satellite OCEANSAT-2 will be launched. This satellite is a follow-on mission to the first one which had a scatterometer and an Ocean Color Monitoring (OCM) instrument onboard both devoted to ocean investigation and applications (prediction of cyclone trajectory, fishery, coastal zone mapping, etc.). For the second mission the Italian Radio Occultation GPS receiver (ROSA) devoted to atmospheric investigation will be added to the two payloads. The ROSA payload was designed and developed for the retrieval of atmospheric pressure, temperature, and humidity profiles. Its performance and specifications will be described in the present paper. The Italian Space Agency (ASI) has promoted the ROSA receiver development to strengthen scientific activities as well as to provide end-user applications. In conclusion the paper highlights the radio occultation related activities in Italy targeted on the next OCEANSAT-2 space mission hosting the ROSA receiver.

Identification of New Road Segments Using a Modified Version of the K-means Algorithm

The present work explains how to identify transformation and new road segments in existing electronic maps by using data coming from devices carried on board on different vehicles, by using a modified version of a standard clustering algorithm called k-means. The working dataset appears as sparse clouds of points with their centres over the road segments, including other different data. Because of the spatial distribution of the points and in order to allow the algorithm to converge in a short number of steps, a simple modification has been implemented. With respect to the standard k-means algorithm which works with a fixed number of clusters, the present modified version works with a large initial number of clusters, with sizes defined a priori on the basis of the smallest possible road segment. The number of clusters is progressively reduced considering, for each steps, only the clusters including a number of points above a specific thresholds. At the end of the algorithm, all the identified clusters are superimposed over a common map in order to validate if a new road segment is identified. The algorithm has been applied on different datasets acquired on the road network of Turin with good results allowing the identification of new road segments not present in the reference map and one-way roads (change in travel direction).

The Italian GPS Radio Occultation experiment on board the Indian satellite oceansat-2

This contribution deals with the description of the Italian Radio Occultation experiment on board the Indian OCEANSAT-2 Mission. Details of the Italian Radio Occultation Ground Segment and results obtained within the validation activity will be discussed. Atmospheric profiles (in terms of refractivity, temperature, and electron density) obtained processing ROSA raw observations will be presented.

Retrieval of atmospheric refractivity profiles from ground-based GPS measurements

An alternative remote sensing technique for the retrieval of atmospheric refractive index profiles is presented. It is an optimization procedure for the inversion of ground-based GPS phase measurements collected at low elevation angles. The Abel inversion, usually adopted to infer atmospheric profiles from space-based radio occultation observations, cannot be applied to the inversion of ground-based measurements. Therefore, we propose a technique which iteratively looks for the refractive index profile by minimizing a cost function. This function depends on several measurements of signal times-of-flight, provided that the correspondent real signal arrival angles could be known with sufficient accuracy. Although different optimization procedures could be used, a non-linear least squares procedure has been adopted in this first attempt. Starting from a different first guess profile, the retrieved refractivity profile is in good agreement with the reference one, which is used to derive the observables.

An example of vertical profile of tropospheric aerosol retrieved from radiance imagery measured by the MOS/A and MOS/B spectrometers over the Mediterranean Sea

The aerosol kernel functions relate, through a Fredholm integral equation of the first kind, the variations in radiance measured by satellites to the variations in aerosol extinction profile, and thus allow profile retrieval from radiance measurements by inverting the set of the radiative transfer equations for different spectral intervals. In the past, the kernel functions were evaluated for the red and near infrared spectral regions both outside and inside molecular absorption bands. In this article they are tailored to the spectral characteristics of MOS/A and MOS/B spectrometers flying onboard IRS-P3. Then, the inversion formulas are applied to real data acquired during a clear-sky day over the Mediterranean Sea to retrieve the variations in the aerosol vertical profile (with respect to the first-guess profile). The agreement between MOS data and the radiance values corresponding to the retrieved profile (through a LOWTRAN simulation) is always better than the agreement between MOS data and the radiances corresponding to a first-guess profile (through a LOWTRAN simulation, too). This gives evidence in favor of the developed retrieval scheme based on kernel functions.

Radars-gauge (dis-)agreement during the 1994 Piedmont flood

A comparison of precipitation amounts as measured by a network of gauges and two different radar systems is presented (an old-generation non-Doppler radar and a new-generation Doppler one). In spite of their different technology, the two radars share very similar conditions, namely: adverse orography, good view of the area of interest from the radar sites (implying severe ground clutter contamination), and large radar-gauge distances. The available radar data consist of maximum reflectivity echoes maps. The agreement between the different measurement systems is characterized in terms of correlation coefficients between radar-gauge data pairs. The analysis confirms that the spatial representativeness of gauge measurements increases with increasing integration period (although limited, of course, by the relatively short flood duration).

Evaluation of aerosol profile variations from satellite measurements through kernel functions for a stratified atmosphere

A linear relationship between the variations in radiance measured in the visible and the near infrared by satellites and those of aerosol extinction profile have been derived, hence reducing the problem to that of solving a Fredholm integral equation of the first kind. Single scattering approximation and a limitation in the aerosol profile variations represent the constraints for the validity of the mathematics procedure. The retrieved profiles, by the linear constrained inversion method, have proved to be accurate in the lower atmosphere, even if simultaneous changes are taking place in the stratosphere. Vice-versa, variations in the stratosphere are retrieved with a lesser degree of accuracy and only as long as no variations occur in the tropospheric haze.