Landrico U. Dalida, Jr.
A Regional Climate Model Version 3 (RegCM3) was used to simulate the seasonal rainfall (June, July, August and September 2001). The model performance was assessed on a smaller domain covering the Philippines using subjective analysis of rainfall pattern and computation of skill scores. The model was tested by shifting the domain 10 degrees to the left of the Control run, ShL, and to the right ShR. ShL covers more landmass than bodies of water while ShR include more ocean than the control run. ShL and ShR were used to determine the effect of changing the boundary position. The model was then tested using different cumulus parameterization schemes: Grell with Fritsch Chappel closure (FC), Grell with Arakawa-Schubert closure (AS) and Kuo. Assessment of the model performance was first based on subjective analyses of the maps in order to determine how well the rainfall patterns were simulated.The model was then assessed according to bias or anomaly, root mean square error and measures of forecast skills: Bias, % Bias Score and Hit rate. The bases for comparison were the monthly rainfall observations of 44 PAGASA stations and Tropical Rainfall Measuring Mission (TRMM). The contingency table used to determine the skill scores were grouped into four threshold levels: 0-100 mm, 100-300 mm, 300-500 mm and >500 mm. Subjective analyses of the rainfall maps reveal that the domain shifted 100 to the left performs better than Con and ShR with ShR a close second. Using areal rainfall instead of point rainfall, shifting the domain reveals that ShL estimates better the rainfall. On the other hand, the choice of cumulus parameterization schemes affects greatly the model's performance. Subjective analyses show that FC generally overpredicts while AS underpredicts. Nevertheless, it is not distinct whether FC and AS have an edge over the other. Regarding the determination of rainfall contribution of convective and non-convective or stable rainfall produced by RegCM to the total rainfall (TPR), the study was able to show that cumulus convection contributed much to the TPR except for some occasion wherein stable rainfall is greater than convective rains. The use of RegCM for simulating seasonal rainfall is promising. The rainfall pattern and the rain-causing synoptic weather systems were simulated or captured in more cases. However, the rainfall amount was found to be highly sensitive to cumulus parameterization. Kuo gives consistently poor performance and is not recommended for further use in RegCM3. More studies using other cumulus parameterization schemes is recommended and compared with FC. To further improve the model performance, other options such as changing some parameters in the biosphere-atmosphere transfer scheme should be investigated.
Esperanza O. Cayanan
Intense southwest monsoon (SWM) rainfall events causing massive landslides and flash floods along the western sections of the Philippines were studied through conduct of observational and numerical analysis. The heavy rainfall during the southwest monsoon season is hypothesized as caused by strong convective activity generated by monsoon westerlies and enhanced by the presence tropical cyclone (TC) in the vicinity of northeastern Luzon. The strong westerlies upon interaction with the mountain ranges along the west coast of Luzon produce strong vertical motion and consequently generate heavy rainfall. Scientific investigation of four SWM rainfall cases were undertaken to prove this hypothesis. The heavy rainfall cases selected are of varying condition in terms of the presence and position of tropical cyclones. The first case involved two TCs within the Philippine area of responsibility (PAR), the second case has a TC crossing Luzon Island, the third case involves a strong typhoon that passed to the NE of Luzon and the fourth case has no tropical cyclone in the PAR. Five year (2002 - 2006) time series of five-day moving average rainfall from eleven western Luzon stations was utilized in the selection of events. Observational analysis revealed that during the rainfall events, a trough is present over the northeastern portion of the Philippines between the islands of Luzon and Taiwan from the surface to 850 hPa level. Heavy rains in the amount of 20 mm/day and greater the recorded for duration of 5 days or longer over the western sections of Luzon from southern station (Sangley) to northern station (Laoag). Due to exposure and topography, the SWM rainfall distribution is maximum over the western part and decreases eastward except for the case of crossing tropical cyclone where maximum rainfall occurred along its path. It was also noted that heavier and longer duration rainfall occurred when tropical cyclones are embedded within the trough. In this situation, the weather condition is characterized by strong southwesterlies/westerlies of more than 20 meters/sec over the South China Sea and western sections of the Philippines from the surface to 850 hPa levels. The trough migrates northward from an initial NW-SE orientation with the axis rotating counterclockwise until it attains a SW-NE orientation with the upper end over southern Japan. As shown in the 850 hPa latitude-time cross section of zonal wind, the migration of the trough is modulated by the 30-60 day northward and 12-24 day westward oscillations. The position of the trough over Luzon Island up to Luzon Strait between Taiwan and Philippines and the strength of the westerlies of 20 meters/sec from the surface to 850 hPa level could serve as indicators of the occurrence of SWM heavy rainfall events. Through the aid of Fouries analysis, the hypothesis on the cause of heavy rainfall during the southwest monsoon has been presented and explained scientifically through the investigation of four selected cases of heavy rainfall events. With the use of Fourier transforms, the total streamflow is decomposed into monsoon basic flow (Wave Number 0 and 1) and tropical cyclone perturbation flow (Wave Numbers 2-23). The procedure was done to isolate the TC and study its effect or contribution to the monsoon rainfall activity. It was shown that the combined westerlies from the basic flow and westerlies generated by the tropical cyclone interact with the Cordillera Mountain ranges along the west coast of Luzon whose peaks are above 2,000 meters. The strong westerlies are forced to rise above the mountains resulting to strong vertical motion that brings about heavy rainfall. The rising motion is enhanced by the convergence of the northwesterlies from the tropical cyclone and the southwesterlies from the basic monsoon flow over the western Luzon. The numerical analysis involved simulation of the four cases of heavy rainfall events using the Fifth Generation Mesoscale Model (MM5) of the National Center for Atmospheric Research (NCAR)/Penn State University (PSU). In the simulation, four convective parameterization schemes (CPS) are tested to see which is best applicable to the local setting. The CPSs tested include Grell, Betts-Miller (BM), Anthes-Kuo (AK) and Kain Fritsch (KF). Generally, MM5 model did not perform well in the daily simulation of the SWM rainfall. The positions of the troughs and the vortices are displaced. However, the 5-day average of windfields and the total rainfall distribution are close to the observed analysis. In the skill test, the threat score and bias score at the heavy rain intensity threshold showed that the MM5 model fairly simulated the total volume of 5-day rainfall. But in terms of quantity of intense events, the rainfall amount is under-estimated. The skill of the model decreases as the rainfall intensity increases which means that the model performed better for less intense rainfall. The model had some difficulty in simulating the intense convection which is also the findings of other studies using MM5. On the most appropriate CPS, Grell was able to capture the major features of the rainfall distribution of the 5-day SWM rainfall events as well as the track of the tropical cyclone that crossed Luzon Island. It performed better than the other schemes as shown by the least mean error and the root-mean square error (RMSE). The diagnostic analysis of the model simulations did not yield encouraging results. The daily simulations of the model are not good enough. It is noted that the isolation of the cyclone waves is not completely accomplished because vortices are still present in the basic flow. The small model output domain could only preserve the cyclone wave vortices but not the large-scale long wave background flow unlike the NCEP analysis which covers the entire globe. Further studies on this regard shall be undertaken in the future.
Marcelino Q. Villafuerte II
This study aims to investigate possible impact of future global warming on rainfall brought by southwest monsoon season in the Philippines. The Climate Type 1 region as classified by Modified Coronas, where the monsoon is well pronounced, served as the pilot area of the study. First, rainfall is investigated historically in terms of its annual and decadal variations. Then, a "nested" modeling is applied to determine if key characteristics can be simulated and find out if these will be altered in the future. The latest version of Abdus Salam International Center for Theoretical Physics (ICPT) Regional Climate Model Version 3 (RegCM3) nested in a coupled atmosphere-ocean General Circulation Model, ECHAM5/MPI-OM, is used. It was run at 40 km horizontal grid spacing covering the entire Philippines. Two periods were analyzed; first 1961-1990 which served as the baseline period and subjected for validation; second, 2010-2039, as the future projection period. Observational datasets which includes surface observation performed by PAGASA; gridded datasets from APHRODITE, Climatic Research Unit (CRU) and ERA 40 were utilized to validate historical model generated rainfall and surface wind. The model performed well in simulating seasonal cycle of rainfall, daily rainfall magnitude/intensity and prevailing surface wind. Spatial details of rainfall were also produced. Projections for the next thirty years revealed an increase in the mean southwest monsoon rainfall by up to 20% primarily during the peak of rainy season. Increase in the mean rainfall might be attributed to more frequent occurrences of days with heavy rain (>50mm) as projected in the future.
Sonny N. Pajarilla
An observational and theoretical study of sea breeze over Samar Island for the month of May 2009, was conducted. The surface observations of PAGASA weather stations in Catarman, Borongan, Guiuan and Catbalogan were utilized in the observational phase of the study. Through surface observations of wind direction and speed, temperature and relative humidity (RH) coupled with FY2D satellite pictures, three characteristic sea breezes have been examined with different prevailing flow, that is southeast, southwest and a variable direction from southeast to southwest. The northern side of the island (Catarman station) consistently manifested the sea breeze characteristics in all three cases. However, a more intense sea breeze was observed with the southeast prevailing flow. The consistent manifestation of sea breeze over Catarman is due to the fact that either southeasterly or southwesterly flow prevailing over the island, the northern side would have an offshore flow of winds which causes the formation of a strong sea breeze circulation. The offshore prevailing wind advects warmer air over land towards the sea which tends to produce a strong horizontal temperature gradient within a thick layer of the atmosphere and a corresponding strong pressure gradient in the surface layers. Contrarily, on the windward side, the prevailing onshore wind advects the colder air over the sea which inhibits the rise in temperature over land. Consequently, the horizontal pressure gradient is weaker and a weaker circulation is developed, yet stronger winds are measured on the windward coast because of the combined effect of the prevailing flow and the weak sea breeze circulation. Hence, Catarman is the most affected area over Samar island of the sea breeze circulation during the month of May where the large scale prevailing flow is southeasterly to southwesterly. The theoretical phase of the study has utilized the fifth-generation PSU/NCAR Mesoscale Model (MM5) a limited-area, nonhydrostatic, terrain-following sigma-coordinate model. Cumulus parameterization sensitivity test has been conducted which resulted to the integration of Kain-Fritsch (KF) into the model runs of MM5. KF had the highest correlation coefficients in terms of temperature and humidity. The model and the observed winds were not in agreement with each other. The model had an over estimated value of wind speed and the direction was not precise. The large error could have been caused by the coarse model domain and human error in conducting manual observations. The modeling results had been consistent with the observed sea breeze characteristics. It was found that the region of positive vertical motion forms first on the leeward coast and moves inland at a rate slower than the total wind. Strong vertical velocities at about 2 mps were observed at 900 mb (~ 1 km) level. The simulated pattern of vertical motion qualitatively agrees with the development of clouds observed by satellites. The depth of the simulated sea breezes agrees with the 2 km sea breeze characteristic of previous studies over tropical regions. It was manifested by the return flow of the model output to be about 850 - 800 mb level or roughly 1.5 - 2 km in height. From a good set of observations and a localized model configuration, the observed sea breezes over Samar island in May 2009 were very well described and simulated. Since the simulations were good, sea breeze characteristics and its associated weather over the area can therefore be forecasted. Higher temporal and spatial resolution of surface observations was recommended to improve observational analyses. Likewise, to reduce the degree of error in the model simulations and future forecasting, a finer resolution and more physics option was also recommended to be integrated in the model runs utilizing high speed computers.
Nievares C. Nivagine
Metro Manila has evidently gone through a transformation from an urban to a highly urbanized region in the country. The interaction of complex wind flow patterns and the effects of urbanization have lead to the changes in mesoscale circulation and, consequently, the time and location of rainfall. A numerical experiment was conducted using MM5 to describe the effects of urbanization on the mesoscale circulation and its associated rainfall. The study of urban effects includes the impact on the heat energy balance, thermodynamic process and circulation. A comparison between observed and simulated temperature, winds and rainfall shows that MM5 is capable to simulated the mesoscale circulation and its associated rainfall. Urban effects are shown to increase temperature over the urbanized area, which is more notable on the boundaries of the urbanized and rural area, including the coastal area, which enhances sea and lake breezes and produces strong convergence zones. The corresponding vertical velocity distribution of convergence zones and occurrence of rain mixing ratio at different atmospheric level shows an increase of rainfall over the northern section of the urbanized area, which is clearly described by the 24 hour accumulated rainfall. In addition, the location of rainfall occurrence is also consistent with the decrease in temperature, sensible and latent heat fluxes.
Sheila S. Schneider
An observational and modeling study of typhoon Reming is conducted. Reming developed from a tropical depression south of Guam. During the next three days, Durian moved westward and intensified rapidly into a super typhoon near the southern tip of Catanduanes Island. Subsequently, Durian continued to move westward through the central Philippines. During this time, the cyclone weakened rapidly. The observational study includes a detailed analysis of the rainfall variations during the passage of the typhoon over the Bicol regions. Rainfall observations from TRMM and rain gauges are used in the analysis. During the time of maximum intensity, the rainfall distribution is characterized primarily by an axially symmetric pattern, with a minimum over the center. A departure from this symmetry consists of a rainfall maximum of about 25mm/hr, which is located southeast of the center. The symmetrical pattern becomes disorganized during the subsequent westward movement of the typhoon. In addition, there is a decrease in the rainfall intensity. The corresponding raingauge observations are generally not consistent with the TRMM observations. This lack of consistency is presumbly due to the highly variable characteristic of natural rainfall variations. In general, the MM5 model simulations of rainfall are not entirely satisfactory, except for the simulations of the 24 hour accumulated rainfall. The relatively poor performance of the MM5 model suggest that further studies should be conducted in order to improve the accuracy of the model.