A cylindrical data set consists of circular and linear variables. We focus on developing an outlier detection procedure
for cylindrical regression model proposed by Johnson and Wehrly (1978) based on the k-nearest neighbour approach.
The procedure is applied based on the residuals where the distance between two residuals is measured by the Euclidean
distance. This procedure can be used to detect single or multiple outliers. Cut-off points of the test statistic are generated
and its performance is then evaluated via simulation. For illustration, we apply the test on the wind data set obtained
from the Malaysian Meteorological Department.
Rainfalls data have been broadly used in researches including in hydrological and meteorological areas. Two common ways in extracting observations from hourly rainfalls data are the window-based analysis (WBA) and storm-event analysis (SEA) approach. However, the differences in the qualitative and quantitative properties of both methods are still vaguely discussed. The aim of studying these dissimilarities is to understand the effects of each approach in modelling and analysis. The qualitative difference is due to the way the two analyses define the accumulated rainfalls for observations which are referred to as rainfall and storm depths, respectively. The repetitiveness of rainfall depths provide nested structure while the storm depths are considered independent. The quantitative comparisons include their statistical and scaling properties that are linked by the self-similarity concept from simple scaling characteristics. If self-similarity concept
holds, then the rainfall or storm depths follow simple scaling and the analysis would be simplified. The rainfall depths showed clearer simple scaling characteristics compared to the storm depths. Though the storm depths do not yield self-similarity for a large range of storm duration but the characteristics of simple scaling can be observed for a reduced range of the considered duration. In general, the context of the research and the region of the time interval and duration will be an important aspects to consider in choosing which method is best to use for analyzing the data.
A hybrid climate model (HCM) is a novel proposed model based on the combination of self-organizing map (SOM) and analog method (AM). The main purpose was to improve the accuracy in rainfall forecasting using HCM. In combination process of HCM, SOM algorithm classifies high dimensional input data to low dimensional of several disjointed clusters in which similar input is grouped. AM searches the future day that has similar property with the day in the past. Consequently, the analog day is mapped to each cluster of SOM to investigate rainfall. In this study, the input data, geopotential height at 850 hPa from the Climate Forecast System Reanalysis (CFSR) are training set data and also the complete rainfall data at 30-meteorological stations from Thai meteorological department (TMD) are observed. To improve capability of rainfall forecasting, three different measures were evaluated. The experimental results showed that the performance of HCM is better than the traditional AM. It is illustrated that the HCM can forecast rainfall proficiently.
Renewable energy assessments for resort islands in the South China Sea were conducted that involves the collection and analysis of meteorological and topographic data. The meteorological data was used to assess the PV, wind and hydropower system potentials on the islands. Furthermore, the reconnaissance study for hydro-potentials were conducted through topographic maps in order to determine the potential sites suitable for development of run-of-river hydropower generation. The stream data was collected for 14 islands in the South China Sea with a total of 51 investigated sites. The data from this study are related to the research article "Optimal combination of solar, wind, micro-hydro and diesel systems based on actual seasonal load profiles for a resort island in the South China Sea" published in Energy (Khan et al., 2015) [1].
This study developed 0.05° × 0.05° land-only datasets of daily maximum and minimum temperatures in the densely populated Central North region of Egypt (CNE) for the period 1981-2017. Existing coarse-resolution datasets were evaluated to find the best dataset for the study area to use as a base of the new datasets. The Climate Prediction Centre (CPC) global temperature dataset was found to be the best. The CPC data were interpolated to a spatial resolution of 0.05° latitude/longitude using linear interpolation technique considering the flat topography of the study area. The robust kernel density distribution mapping method was used to correct the bias using observations, and WorldClim v.2 temperature climatology was used to adjust the spatial variability in temperature. The validation of CNE datasets using probability density function skill score and hot and cold extremes tail skill scores showed remarkable improvement in replicating the spatial and temporal variability in observed temperature. Because CNE datasets are the best available high-resolution estimate of daily temperatures, they will be beneficial for climatic and hydrological studies.
Investigation of meteorological disasters caused by small-scale topography shows that flashover due to wind age yaw occurred quite often where col topography existed. Considering that the distribution pattern of wind profile at different locations of a col topography is not clear, this paper, with wind tunnel tests, studied the influenced of such topographic features of a col as hill slope and valley mouth width on the wind profile at different locations. The results of wind tunnel tests indicated that over-hill wind has a stronger effect on wind velocity correction coefficient than does valley wind, that compared to flat terrain wind velocity, the maximum speed-up amplitude of wind velocity at valley throat and hill summit reach 33 and 53%, respectively, apparently higher than 10% specified in Codes, that wind velocity at valley throat increases with the increase of hill slope and decreases with the increase of valley mouth width, that wind velocity in the valley basically does not go up when the slope of one hill side is smaller than 0.1 and that wind velocity at the same non-dimensional height of a 3D hill summit increases with the increase of hill height.
The prediction of rainfall on monthly and seasonal time scales is not only scientifically challenging but is also important for planning and devising agricultural strategies. In this paper, the study is conducted to examine the pattern of monthly rainfall in Alor Setar, Kedah within ten years which is from 2008 to 2018. This paper considered a model based on real data that obtained from Department of Meteorology Malaysia. This study indicates that the monthly rainfall in Alor Setar has a seasonal and trend pattern based on yt vs t plotting, autocorrelation function and Kruskal Wallis Test for seasonality. The examined rainfall time-series modelling approaches include Naïve Model, Decomposition Method, Holt-Winter’s and Box-Jenkins ARIMA. Multiplicative Decomposition Method was identified as the best model to forecast rainfall for the year of 2019 by analysing the previous ten-year’s data (2008-2018).As a result from the forecast of 2019, October is the wettest month with highest forecasted rainfall of 276.15mm while the driest month is in February with lowest forecasted rainfall of 50.55mm. The model is therefore adequate and appropriate to forecast future monthly rainfall values in the catchment which can help farmers to plan their farming activities ahead of time.
The low resolution Automatic Picture Transmission (APT) data from the National Oceanic and Atmospheric Administration (NOAA) polar-orbiting satellites Advanced Very High Resolution Radiometer (AVHRR) is being received and recorded in real-time mode at ground receiving station in School of Science and Technology, Universiti Malaysia Sabah. The system is suitable for the developing and undeveloped countries in south and Southeast Asia and is said to be acceptable for engineering, agricultural, climatological and environmental applications. The system comprises a personal computer attached with a small APT receiver. The data transmission between the ground receiving station and NOAA satellites is using the electromagnetic wave. The relation for receiving and processing the electromagnetic wave in the transmission will be discussed.
Spatio-temporal datasets are a collection of datasets where data can vary in both space and time. Theoretically, such datasets can be considered as continuous and discrete. For example, specification of the function, F: Ed T Rn, where Ed denotes d-dimensional Euclidean space, T = R* ∩ {} the domain of time and Rn an n-dimensional scalar field. Examples of such data sets include time-varying simulation results, film and videos, time-varying medical datasets, geometry models with motion or deformation, meteorological measurements, and many more. It is therefore highly desirable to use visualisation to summarize meaningful information in higher dimensional spatio-temporal datasets. Our aim is to conceive an efficient visual study to facilitate scientists in identifying temporal association among complex and chaotic atom movements in ion trajectories. An application that uses a streamline for spatial motion of ion trajectories and Colour Number Coding Scheme for temporal encoding of high degree of timeline events among mobile ions is proposed. With an anthology of the visual examples, it was revealed that this application would be beneficial for scientists to visually mine any 3D spatio-temporal dataset.
The analysis of the spatial data has been carried out in many disciplines such as demography, meteorology, geology and remote sensing. The spatial data modelling is important because it recognizes the phenomenon of spatial correlation in field experiments. Three main categories of the spatial models, namely, the simultaneous autoregressive (SAR) models (Whittle, 1954), the conditional autoregressive (CAR) models (Bartlett, 1971), and the moving average (MA) models (Haining, 1978) have been studied. Whittle (1954) presented a form of bilateral autoregressive (AR) models, whereas Basu and Reinsel (1993) considered the first-order autoregressive moving average (ARMA) model of the quadrant type. Awang, N. and Mahendran Shitan (2003) presented the second-order ARMA model, and established some explicit stationary conditions for the model. When fitting the spatial models and making prediction, it is assumed that, the properties of the process would not change with sites. Properties like stationarities have to be assumed, and for this reason, it was therefore imperative that the researchers had made certain that the process was stationary. This could be achieved by providing the explicit stationarity conditions for the model. The explicit conditions, for a stationary representation of the second-order spatial unilateral ARMA model denoted as ARMA(2,1;2,1), have been established (Awang, N. and Mahendran Shitan, 2003) and in this paper, some explicit conditions are established for a stationary representation of the second-order spatial unilateral ARMA model, denoted as ARMA(2,2;2,2).
Simulation is used to measure the robustness and the efficiency of the forecasting
techniques performance over complex systems. A method for simulating multivariate
time series was presented in this study using vector autoregressive base-process. By
applying the methodology to the multivariable meteorological time series, a simulation
study was carried out to check for the model performance. MAPE and MAE performance
measurements were used and the results show that the proposed method that consider
persistency in volatility gives better performance and the accuracy error is six time smaller
than the normal hybrid model.
This study highlights the advantage of functional data approach in assessing and comparing the PM10 pollutant behaviour as an alternative statistical approach during and between the two extreme haze years (1997 and 2005) that have been reported in Selangor, state of Malaysia. The aim of the study was to improvise the current conventional methods used in air quality assessment so that any unforeseen implicit information can be revealed and the previous research findings can be justified. An analysis based on the daily diurnal curves in place of discrete point values was performed. The
analysis results provided evidences of the influence of the change in the climate (due to the El-Nino event), the different levels of different emission sources and meteorological conditions on the severity of the PM10 problem. By means of the cummulative exceedence index and the functional depth method, most of the monitoring stations for the year 2005 experienced the worst day of critical exceedences on the 10th of August, while for the year 1997 it occurred between 13th and 26th September inclusively at different dates among the stations.
Estimation and forecast of groundwater recharge and capacity of aquifer are essential issues in water resources investigation. In the current research, groundwater recharge, recharge coefficient and effective rainfall were determined through a case study using empirical methods applicable to the tropical zones. The related climatological data between January 2000 and December 2010 were collected in Selangor, Malaysia. The results showed that groundwater recharge was326.39 mm per year, effective precipitation was 1807.97 mm per year and recharge coefficient was 18% for the study area. In summary, the precipitation converted to recharge, surface runoff and evapotranspiration are 12, 32 and 56% of rainfall, respectively. Correlation between climatic parameters and groundwater recharge showed positive and negative relationships. The highest correlation was found between precipitation and recharge. Linear multiple regressions between
recharge and measured climatologic data proved significant relationship between recharge and rainfall and wind speed. It was also proven that the proposed model provided an accurate estimation for similar projects.
Accurate prediction of any type of natural hazard is a challenging task. Of all the various hazards, drought prediction is challenging as it lacks a universal definition and is getting adverse with climate change impacting drought events both spatially and temporally. The problem becomes more complex as drought occurrence is dependent on a multitude of factors ranging from hydro-meteorological to climatic variables. A paradigm shift happened in this field when it was found that the inclusion of climatic variables in the data-driven prediction model improves the accuracy. However, this understanding has been primarily using statistical metrics used to measure the model accuracy. The present work tries to explore this finding using an explainable artificial intelligence (XAI) model. The explainable deep learning model development and comparative analysis were performed using known understandings drawn from physical-based models. The work also tries to explore how the model achieves specific results at different spatio-temporal intervals, enabling us to understand the local interactions among the predictors for different drought conditions and drought periods. The drought index used in the study is Standard Precipitation Index (SPI) at 12 month scales applied for five different regions in New South Wales, Australia, with the explainable algorithm being SHapley Additive exPlanations (SHAP). The conclusions drawn from SHAP plots depict the importance of climatic variables at a monthly scale and varying ranges of annual scale. We observe that the results obtained from SHAP align with the physical model interpretations, thus suggesting the need to add climatic variables as predictors in the prediction model.
Drought affects vegetation growth to a large extent. Understanding the dynamic changes of vegetation during drought is of great significance for agricultural and ecological management and climate change adaptation. The relations between vegetation and drought have been widely investigated, but how vegetation loss and restoration in response to drought remains unclear. Using the standardized precipitation evapotranspiration index (SPEI) and the normalized difference vegetation index (NDVI) data, this study developed an evaluation framework for exploring the responses of vegetation loss and recovery to meteorological drought, and applied it to the humid subtropical Pearl River basin (PRB) in southern China for estimating the loss and recovery of three vegetation types (forest, grassland, cropland) during drought using the observed NDVI changes. Results indicate that vegetation is more sensitive to drought in high-elevation areas (lag time 8 months). Vegetation loss (especially in cropland) is found to be more sensitive to drought duration than drought severity and peak. No obvious linear relationship between drought intensity and the extent of vegetation loss is found. Regardless of the intensity, drought can cause the largest probability of mild loss of vegetation, followed by moderate loss, and the least probability of severe loss. Large spatial variability in the probability of vegetation loss and recovery time is found over the study domain, with a higher probability (up to 50 %) of drought-induced vegetation loss and a longer recovery time (>7 months) mostly in the high-elevation areas. Further analysis suggests that forest shows higher but cropland shows lower drought resistance than other vegetation types, and grassland requires a shorter recovery time (4.2-month) after loss than forest (5.1-month) and cropland (4.8-month).
This study applies the clonal selection algorithm (CSA) in an artificial immune system (AIS) as an alternative method to predicting future rainfall data. The stochastic and the artificial neural network techniques are commonly used in hydrology. However, in this study a novel technique for forecasting rainfall was established. Results from this study have proven that the theory of biological immune systems could be technically applied to time series data. Biological immune systems are nonlinear and chaotic in nature similar to the daily rainfall data. This study discovered that the proposed CSA was able to predict the daily rainfall data with an accuracy of 90% during the model training stage. In the testing stage, the results showed that an accuracy between the actual and the generated data was within the range of 75 to 92%. Thus, the CSA approach shows a new method in rainfall data prediction.
The incident involving hydrogen release in industry has become a major concern since numerous incidents were observed to have occured over the years. This paper is designed to do the consequence modelling and analysis using PHAST Simulator for the release rate, potential fire and vulnerability to human by lethality versus probit simulated at 5 mm, 35 mm and 70 mm leak scenarios and three types of atmospheric stability at hydrogenation plant in Telok Panglima Garang. The simulation was carried out by inputting data of leak scenario, meteorological data, material data and process data related to the hydrogenation plant. The simulation results were analyzed and discussed on the discharge rate, dispersion concentration and effect of jet fire such as flame length, downwind distance and lethality for radiation intensity level of 4 kW/m2 , 12.5 kW/m2 and 37.5 kW/m2 . Based on the results, the discharge rate and radiation intensity are dependent on the leak sizes regardless of the different atmospheric conditions. However, the dispersion is dependent on both atmospheric stability and leak sizes. Lastly, the lethality and area of impact are simulated from the radiation intensity produced by the jet fire for each leak size. To conclude, adoption of PHAST software is vital for consequence modelling as this software is able to illustrate the outcomes of hazards due to loss of containment and with this will enable related personnel to respond effectively to any hazardous incidents. As a recommendation, hydrogen fixed gas detectors are proposed for installations at specific location after taking into account the smallest leak that may happen which is at 5 mm leak size.
The vector autoregressive (VAR) approach is useful in many situations involving model development for multivariables
time series. VAR model was utilised in this study and applied in modelling and forecasting four meteorological variables.
The variables are n rainfall data, humidity, wind speed and temperature. However, the model failed to address the
heteroscedasticity problem found in the variables, as such, multivariate GARCH, namely, dynamic conditional correlation
(DCC) was incorporated in the VAR model to confiscate the problem of heteroscedasticity. The results showed that the use
of the VAR coupled with the recognition of time-varying variances DCC produced good forecasts over long forecasting
horizons as compared with VAR model alone.
A noticeable increase in drought frequency and severity has been observed across the globe due to climate change, which attracted scientists in development of drought prediction models for mitigation of impacts. Droughts are usually monitored using drought indices (DIs), most of which are probabilistic and therefore, highly stochastic and non-linear. The current research investigated the capability of different versions of relatively well-explored machine learning (ML) models including random forest (RF), minimum probability machine regression (MPMR), M5 Tree (M5tree), extreme learning machine (ELM) and online sequential-ELM (OSELM) in predicting the most widely used DI known as standardized precipitation index (SPI) at multiple month horizons (i.e., 1, 3, 6 and 12). Models were developed using monthly rainfall data for the period of 1949-2013 at four meteorological stations namely, Barisal, Bogra, Faridpur and Mymensingh, each representing a geographical region of Bangladesh which frequently experiences droughts. The model inputs were decided based on correlation statistics and the prediction capability was evaluated using several statistical metrics including mean square error (MSE), root mean square error (RMSE), mean absolute error (MAE), correlation coefficient (R), Willmott's Index of agreement (WI), Nash Sutcliffe efficiency (NSE), and Legates and McCabe Index (LM). The results revealed that the proposed models are reliable and robust in predicting droughts in the region. Comparison of the models revealed ELM as the best model in forecasting droughts with minimal RMSE in the range of 0.07-0.85, 0.08-0.76, 0.062-0.80 and 0.042-0.605 for Barisal, Bogra, Faridpur and Mymensingh, respectively for all the SPI scales except one-month SPI for which the RF showed the best performance with minimal RMSE of 0.57, 0.45, 0.59 and 0.42, respectively.
General flowering (GF) is a unique phenomenon wherein, at irregular intervals, taxonomically diverse trees in Southeast Asian dipterocarp forests synchronize their reproduction at the community level. Triggers of GF, including drought and low minimum temperatures a few months previously has been limitedly observed across large regional scales due to lack of meteorological stations. Here, we aim to identify the climatic conditions that trigger large-scale GF in Peninsular Malaysia using satellite sensors, Tropical Rainfall Measuring Mission (TRMM) and Moderate Resolution Imaging Spectroradiometer (MODIS), to evaluate the climatic conditions of focal forests. We observed antecedent drought, low temperature and high photosynthetic radiation conditions before large-scale GF events, suggesting that large-scale GF events could be triggered by these factors. In contrast, we found higher-magnitude GF in forests where lower precipitation preceded large-scale GF events. GF magnitude was also negatively influenced by land surface temperature (LST) for a large-scale GF event. Therefore, we suggest that spatial extent of drought may be related to that of GF forests, and that the spatial pattern of LST may be related to that of GF occurrence. With significant new findings and other results that were consistent with previous research we clarified complicated environmental correlates with the GF phenomenon.