Tropical Rain/WIPAS Project

G.D. Abrajano, R.H.N. Ching, J.L.F. Zamora, H.C. Dy, A.S. Gimpaya Jr., J.C.A. Pusta, J.R.M. San Luis, E.M.S. Trono, S. Gonzales, A. Lotho, N. Libatique, C. Pineda and G. Tangonan (also please check out Wipas Project page)

Abstract

This study focuses on the behavior of microwave links during rain events, especially tropical rains. Our research uses acoustic sensors/recorders which are placed inside tin cans, specifically 1/2 gallon ice cream cans. We do recordings for two cases: (1) short distance recordings where cans with recorders are placed in different areas of the Ateneo campus (i.e. Manila Observatory, Sec A, CTC, football field, Moro Lorenzo), and (2) long distance recordings done at night when we are at our homes, where we leave the acoustic sensors on our rooftops. For the first case, the distance between recorders is a few hundred meters, while for the second case the distance is a few kilometers.

To adhere to international standards for rain measuring and monitoring, we use tipping buckets located at the Manila Observatory and PLDT Balara in order to measure how strong the rain is on a specific time interval. From this we can get the specific amount of rainfall, and relate this with the power spectrum of the rain recordings that we have.

Significance of the Study

The behavior of tropical rains can be very different from that of rains in other regions of the world. The Philippines in particular experiences several heavy rainfalls during the wet season, while some parts of the country experience rainfall all throughout the year. Tropical rains can be very dynamic, some lasting for several hours while others lasting for only ten minutes. Rain cell sizes also vary from one rain event to another, sometimes taking up a whole region, while most of the time only taking up a particular area in the region. Because of this burst-like behavior and varying strength of tropical rains, it is very important to somehow find a way to characterize them. Obviously, standard rain monitoring tools like the tipping buckets pose some limitations in completely characterizing tropical rains. A per-minute data logging system of a tipping bucket would not be able to give us much information for a ten minute rain. Because of this, an alternative rain monitoring system is being studied in this research.

Studying rain events can have several applications. Rain characteristics are important for several fields, including climate studies, disaster prevention, and communications systems research. A particular application of this year’s rain thesis is a correlation of microwave link attenuation to the strength of the rain and the size of the rain cell. The equipment being tested is the 26-GHz Wireless Internet Protocol Access System or WIPAS donated by the Japan Radio Company. We would like to determine the system performance when the technology is subjected to the types of rain that we have here in the Philippines.

Summary of Previous Research Done

SY 2006-2007

This group was able to measure tropical rain in real time using digital acoustic recorders. They were already able to establish the protocols for rain monitoring and they already have very significant rain data. They were able to show from the data how tropical rain is fundamentally burst-like in nature and how the peak rain rate can be much larger than the average rain rate. These are data that cannot be readily acquired from time-averaged results like that from a tipping bucket. They also saw that for distances of less than 8 kilometers there was an average burst length of 15 seconds. This shows the spatial and temporal behavior of rain. However, for distances more than 8 kilometers, the results show that not all locations receive the same quality of rain despite it occurring at the same time.

SY 2007-2008

We started working last summer, setting up and studying protocols for rain monitoring. On the early weeks of the semester, we adopted last year’s monitoring protocol and settings. However, we soon found out that the settings are too high for these season’s rains/typhoons so we conducted some calibration experiments. We also acquired more recorders so settings had to be studied and revised.

Basically, for rain monitoring, we set up the acoustic recorders at several points on the campus if the rain event is in the morning. When the rain event happens at night, the members of the tropical rain would start recording in their own houses. This way we can do short distance monitoring during the day and long distance monitoring at night. Along with the acoustic sensors, we also had tipping buckets stationed at the Manila Observatory, CTC Rooftop, and PLDT-Balara to adhere to international standards for rain monitoring.

In parallel with the rain monitoring efforts, the group has also been maintaining the WIPAS link. A short-distance link was setup during the summer between MVP-CSL Building and the Blue Eagle Gym, the distance being 400 meters. At around September the long-distance link was established with the help of JRC engineers between the MVP-CSL Building and PLDT Balara Exchange at Commonwealth Ave., Quezon City. The link distance is approximately 4 kilometers. Both stations have desktop computers connected to them where several transmission parameters like transmission level, receiving level, and Bit-Error Rates are being logged.

Significant Results/Data

For the short-distance link between MVP-CSL Building and Blue Eagle Gym, we got a significant rain event last July 5, 2007. From the logs on the computer, we observed significant attenuation of the transmitted signal. Fig. 1 shows the receiving level of the WIPAS terminals, while Fig. 2 shows the tipping bucket data for the same time interval.

• Fig.1:
  

• Fig.2:
  

From these two figures we can clearly see that the peak signal attenuation happened during the peak rain rate. Since the rain event lasted for some time, it did not really matter that the tipping bucket is logging a per minute rain rate. However, since the transmission logger is logging in every second, we can see a more detailed rain behavior from the acoustic sensor data. The peak ten-minute rain is shown in Fig.3.

• Fig.3:
  

For the long-distance link, we had a rain event last October 2, 2007. The WIPAS equipments were running at that time, and we recorded the rain from several points along the link – Faura Building, SEC-A, Manila Observatory, Moro Lorenzo Sports Complex, Capitol Hills Clubhouse, and PLDT-Balara. We were basically able to cover the whole 4 kilometer link, and we were able to see some significant attenuation. The receiving level is shown in Fig.4. The acoustic rain data gathered that time are shown in Fig.5.

• Fig.4:
  

• Fig.5:
  

We can see from Fig.5 that the rain rates are different for the different sites. The rain is very heavy at the Ateneo side while at the PLDT-Balara side it was just a light rain. Even from this short distance we were able to show how the rain strength varies. We also observed that for this particular rain event, the link went down, or there was enough attenuation for the link to lose the connection. However, after some time the link went back up again.

WIPAS Update as of February 22, 2008

Downtimes Due to Precipitation

There are three criteria used to determine the downtimes. The first two are based on the Received Signal Level (RSL) while the third is based on the Bit Error Rate (BER). The transmitted signal has been considerably attenuated when the RSL reached -80dB and lower, and when it reached -99dB the link is completely down. As for the BER criterion, the accepted BER level is lower than 10^-6, and this also corresponds to the lower RSL (-80dB and lower). Dates with downtimes

SY 2008-2009 Updates

Edited by Hansel Dy.

The new additions to the team are Dy, Gimpaya, Pusta, San Luis, and Trono, all supersenior undergraduates, and also Alfred Lotho and Silver Gonzales, senior Computer Science majors.

The group is working on designing and implementing a low-cost sound detector circuit that will be used in determining the intensity of rainfall. The device consists of a speaker, an audio transformer and a bridge rectifier. The speaker generates small AC voltages from the vibrations (sounds) it detects. These voltages, in the range of a hundred µvolts are then passed through cascading audio amplifiers that magnify them to values that can pass through the bridge rectifier. The different voltage levels that are generated by the sounds can be observed by obtaining the voltage readings across the ceramic capacitor in the bridge. The voltage levels attained by the charging capacitor indicate how loud the detected sounds are hence indicating the intensity of the rain.

Shown below is a schematic of the device:

• schematic.JPG:
  

• The updated "low-cost rain sensor":
  *

• Figure 1. Accumulated Rain and WIPAS’ Received Signal Level (RSL) vs. Time (08 September 2008, 18:00 – 24:00):
  

The above figure shows the correlation of the rain intensity and its effects on the WIPAS link’s RSL during a rain event induced by Typhoon Marce (International Codename Sinlaku). Notice that at around 18:30-19:10 local time, the RSL dropped significantly to -99 dBm, while the rain accumulated from the Casella tipping bucket climbed quickly to 12 mm. But after about 5 hours, the RSL level returned to an average of -72 dBm, while the rain accumulated for that period was only around 8 mm, for a total of 20.8 mm accumulated rain over that Monday night.

Please go to the lab notebooks for updates...