Gale Warning
Issued at: 05:00 AM TODAY , 23 January 2022
Gale Warning # 25 (Final)

Gale Warning in PDF file

Weather Advisory
Issued at: 11:00 PM 2022 January 19
Weather Advisory in PDF file

General Flood Advisories - Regional
 General Flood Advisory issued as of 7 PM, 20 August 2018


Monthly Climate Assessment and Outlook (July-August 2018)

Issued: 06 August 2018

Monthly Rainfall Forecast
RAINFALL FORECAST  (September 2018 - February 2019) 
UPDATED: 29 August 2018 (next update September 26, 2018)

Regional Rainfall Forecast
Issued: 29 August 2018
Valid for: September 2018 - February 2019
Farm Weather Forecast and Advisories
ISSUED: 8 AM,  FRIDAY,  MAY 24, 2019
, 2019
FWFA:  N0. 19-144

Ten-Day Regional Agri-Weather Information
DEKAD NO. 15   MAY 21 - 31, 2019
The weather systems that will affect the whole country are the frontal system, easterlies, ridge of high pressure area, intertropical convergence zone (ITCZ) and low pressure area (LPA).

Seasonal Climate Outlook
Issued:  13 July 2018
FOR July - December 2018

Astronomical Diary
Issue for October 2018
The October Orionids meteor shower will be active from October 17-25, 2017.

   PAGASA webmail  Learning Tools  Bidding 

  Training Activities  Vacancy  Research-and-Development

  Engineering  faqupdate  Publications 

Aida M. Jose

Semestral and 12-month rainfall anomalies subjected to power spectrum analysis using the approach of Blackman and Tukey. Analysis revealed 2 significant quasi-periodicities with wavelengths equivalent to 2 to 2.5 years and 4 to 5 years. These quasi-periodicities which possibly represent certain pulsation in the general atmospheric circulation were considered in the choice of plausible predictors of seasonal rainfall performance. Empirical equations to predict seasonal rainfall performance were then developed for various stations in the Philippines using multiple linear regression. The choice of plausible predictors were based on diagnostics of the physical mechanisms of abnormalities in the seasonal evolution of large-scale circulation patterns affecting the climate of the Philippines. Five categories of potential predictors used in the regression are composed of (a) pressure tendency (b) seasonal pressure (c) pressure gradient (d) monthly pressure (e) surface air temperature and (f) upper-air mean zonal wind. The results of the regression analysis indicated that rainfall anomalies at 37 stations in the Philippines responded significantly and selectively to particular predictors. There are certain areas in which the seasonal rainfall are more predictable in a particular semester than the other. Comparison of the predicted values based on independent data sets, with the observed values for 3 sample stations showed that the predicted and the observed values are positively and closely related to each other. Because of the diversity of seasonal rainfall variation and its responses to various predictors from station to station, comparisons of the spatial distribution of the predicted and observed values for 1960 and 1980 were made. Specific areas of predicted categories of seasonal rainfall anomalies were identifiable in the spatial distribution patterns for the observed values. Scatter diagrams comparing the predicted and observed values at various points of observations during the 2 semesters of both years showed remarkable goodness of fit. Examination of the yearly values of the regression coefficients for 3 sample stations indicated certain degree of stability. However, for operational purposes, updating of such coefficients should be made as a component of the scheme to assure that the behaviour of the predictors be within the atmospheric mode of the most recent past. Isolated discrepancies in the prediction results could be attributed mainly to either too slow or abrupt change in the seasonal evolution of the general circulation patterns which can not be captured by the regression scheme. In addition to this limitation, is the inherent disadvantage in the use of regression which tends to underestimate predicted values from that of the observed.

Raquel V. Francisco

A two-dimensional, slab-symmetric cloud model with detailed microphysics is presented. Drops were classified into 37 size classes representing a droplet spectrum from 1 um to 4 mm radius. Each class of droplet is subjected to condensation and collection processes. New droplets were formed by the nucleation process. The sensitivity of the model to changes in the eddy diffusion coefficients is discussed. The effect of the magnitude of initial perturbation, environmental winds, and cloud interactions on cloud growth are also investigated. Results show that the model is capable of simulating the life history of a warm cloud from initiation to dissipation stage. Comparison with actual observations based on day 261 of GATE reveal some realistic features of the model. Numerical experiments on simultaneous clouds show that the upshear cloud grows stronger than its downshear counterpart. Experiments on non-simultaneous clouds reveal that the younger cell grows at the expense of the older cell.

Prisco D. Nilo

Meteorological variables which are likely associated with Zamboanga wet season rainfall and onset date are examined in a stepwise regression analysis. Consequently, a long-range forecasting scheme for Zamboanga wet reason rainfall and onset date anomaly is developed. The regression equations involved two predictors each for rainfall and onset date. Three predictors include Darwin sea level pressure and Jakarta rainfall. The forecast experiments for the period 1947-83 show that the regression models for Zamboanga wet season rainfall and onset date have a considerable accuracy in predicting extreme anomalies. However, the results also show that the forecasts for normal conditions as well as the overall accuracy is not good. The statistical method using geographical technique was also examined, and the results of the tests show that graphical approach is inferior to the use of regression models. The various aspects of regression and graphical forecasting scheme are discussed.

Imelda I. Valeroso

A method for predicting the movement of tropical cyclones based on Model Output Statistics is presented. Tropical cyclones which entered the Philippine Area of Responsibility (PAR) for the period covering 1978-1986 consisting of 557 forecast cases were utilized as the development data set for the experiment. Latitude-longitude positions of tropical cyclones at initial time, and the forecasts of the latest modified, operational Barotropic Model of PAGASA using a 2x2 grid system, were taken as predictors and the predictands included the 12-hourly forecasts. The simple linear and the multi-linear regression equations from both the unstratified and the stratified data, were utilized as predictive equations and were the bases of the new Model Output Statistics forecast. Seven (7) tropical cyclones consisting of 57 forecast cases which occurred in 1987 and 1988 were used as the independent data for forecast verification. Results show the usefullness of statistical correction in improving the forecast of the operational Barotropic Model. The latitude forecast improvement of the operational Barotropic Model forecast by the use of the "direct" MOS scheme, whereby the regression equations obtained from the forecasts of the model were utilized at the predictive equations is remarkable. The over-all latitude forecast improved by 40.70% at the 24-hr forecast period, and by 41.95% for the entire 72-hr forecast, period, by the application of the multilinear quarterly regression equation. This is followed by the multi-linear monthly (32.13% over-all 24-hr forecast), and the simple linear quarterly (36.52% entire 72-hr forecast).

Shirley V. Almazan, Catalino P. Arafiles, Bernardo M. Soriano, Jr.

The Manila Bay is one of the Philippines' important bodies of water which is being subjected to tremendous stresses due to the burgeoning population of the Metropolitan Manila Area. Several human activities such as reclamation and wastewater disposal have greatly affected the features of the Bay. These activities and their impacts on the coastal environment are presented in this paper.

Carina G. Lao

The long period variations of tropical cyclone occurrence (both frequency and tracks) have been studied. These variations of tropical cyclone have been related with rainfall, pressure, El Niño and the Quasi-Biennial Oscillation. The resulting relationships have been used to develop statistical methods for seasonal prediction.

The frequency of the tropical cyclone occurrence is specified in terms of cyclone days. The relationship between these cyclone days and the different variables mentioned earlier are studied by the ranking method, correlation analysis, compositing and spectral analysis. Prediction equations are developed to predict the total annual number of cyclone days, as well as the corresponding number for selected periods of the typhoon season.

The significant periods which are revealed by the analysis of the long period variations of the cyclone days correspond to periodicities of 36, 24 and 14 months. These periods are also found in the time series of rainfall and pressure. The periodicity at 24 months could be related to the Quasi-Biennial Oscillation.

The study also indicates that the most important relationship is between the frequency of tropical cyclone occurrence and rainfall in Nauru and Galapagos; and pressure in Darwin, which occur a few months before the typhoon season.

With respect to El Niño, there are more tropical cyclones during periods of moderate to strong El Niño. On the otherhand, with respect to Quasi-Biennial Oscillation, more tropical cyclones occur when the 30 mb. zonal wind has easterly component.

The regression equations which have been developed in this study are tested using dependent and independent data sets. The results show that the most accurate prediction equation is the one which predicts the number of cyclone days for the second half of the cyclone season (SEPD).

Eugenio M. Aquino, Esperanza O. Cayanan, Cynthia P. Celebre

A graphical method of forecasting rainfall over northeastern Mindanao was developed. Eighteen weather elements observed at NAHA (Japan) at 0000Z for a period of 20 years for the months of December, January and February were considered. Four out of these elements were used as predictors. Rainfall data from three stations in the Philippines were utilized as the dependent variable in the study. The method showed encouraging results.