Burapha Science Journal

Influence of Water Quality on the Community Structure and Diversity of Phytoplankton in Shellfish Farming Areas, Chonburi Province, Thailand

Mon, 11 May 2026 00:00:00 +0700

Background and Objectives: Chonburi Province is a key coastal area along the upper Gulf of Thailand with high economic and social importance, particularly for marine shellfish farming, which represents a major coastal fisheries activity in eastern Thailand. This activity plays an essential role in supporting coastal livelihoods and contributing to national food security. Shellfish farming systems in this region rely heavily on natural environmental conditions, especially water quality and phytoplankton communities, which function both as the primary food source for shellfish and as primary producers in coastal ecosystems. Phytoplankton species composition, abundance, and diversity are highly sensitive to variations in physical and chemical water parameters, including temperature, salinity, and dissolved inorganic nutrients. Alterations in phytoplankton community structure, particularly blooms of potentially harmful species, may negatively affect water quality, lead to toxin accumulation in shellfish, and pose risks to food safety. Therefore, understanding the interactions between water quality and phytoplankton dynamics is crucial for sustainable management of shellfish farming areas. This study aimed to assess the relationships between water quality parameters and the structure and diversity of phytoplankton communities in shellfish farming areas along the coast of Chonburi Province. Temporal variations in phytoplankton species composition were also examined to support monitoring of potentially harmful phytoplankton and to provide scientific information for effective and sustainable shellfish aquaculture management.

Methodology: Field surveys were conducted in coastal waters of Chonburi Province, Thailand, adjacent to shellfish farming areas. A total of 15 sampling stations were established in Mueang District, with sampling undertaken during five periods between December 2023 and August 2024 to capture seasonal variability. In situ measurements of temperature, salinity, dissolved oxygen (DO), and pH were performed using a multiparameter probe. Surface water samples were collected at approximately 30 cm depth in triplicate at each station, stored in acid-cleaned HDPE bottles, and maintained at 4°C prior to laboratory analysis within 24 hr. Samples for dissolved inorganic nutrients were filtered through pre-combusted glass fiber filters (GF/F) and stored at 4°C, whereas chlorophyll-a samples were stored at -20°C until analysis. Phytoplankton samples were obtained by filtering 20 L of seawater through a 20-µm mesh plankton net and preserved with 4% formalin. Taxonomic identification was performed to the lowest possible level using light microscopy, and abundance was quantified with a Sedgewick–Rafter counting chamber.Community structure was assessed using the Shannon–Wiener diversity index (H’) and Pielou’s evenness (J’). Spatial and temporal variations were evaluated using one-way analysis of variance (ANOVA) followed by Tukey’s HSD test. Relationships between environmental variables and phytoplankton assemblages were examined using Pearson’s correlation and multiple linear regression, with model performance assessed by the coefficient of determination (R²). All statistical analyses were performed using Minitab Statistical Software.

Main Results: Water quality in shellfish farming areas along the coastal zone of Chonburi Province exhibited pronounced seasonal variability. Seawater temperature ranged from 29.2 to 31.3 °C and showed no significant spatial differences among stations, whereas salinity, dissolved oxygen (DO), chlorophyll-a, and dissolved inorganic nutrients varied significantly over time (p < 0.05). Salinity decreased during the rainy season due to freshwater input, while higher and more stable salinity in the dry season favored phytoplankton growth. Chlorophyll-a concentrations ranged from 4.4 to 14.4 µg L^-1, with peak values observed during the dry season, indicating elevated phytoplankton biomass. Phytoplankton assemblages comprised three divisions—Cyanophyta, Chlorophyta, and Chromophyta with a total of 68 genera. Chromophyta dominated throughout the study and exhibited pronounced blooms during the dry season, resulting in the highest phytoplankton density in February (mean: 527,504 cells L^-1; maximum: 1,337,125 cells L^-1), whereas densities declined markedly during the rainy season. Phytoplankton diversity (H’) and evenness (J’) decreased during diatom (Bacillariophyceae) bloom events and increased toward the late rainy season, reflecting a more heterogeneous community structure. Potentially harmful dinoflagellates were detected, particularly during the dry season, including Prorocentrum spp. (6,500 cells L^-1) and Gymnodinium sp. (800 cells L^-1). Correlation and multiple linear regression analyses identified salinity, temperature, and dissolved inorganic nutrients as key drivers regulating phytoplankton density. These findings highlight the seasonal risk of phytoplankton blooms and harmful species, with potential implications for water quality and shellfish safety in the study area.

Conclusions: Overall, the results demonstrated that temperature, salinity, and dissolved inorganic nutrients were key drivers influencing phytoplankton density, diversity, and community composition in the study area. During the dry season, phytoplankton blooms dominated by Bacillariophyceae (diatoms) were observed, resulting in reduced diversity, whereas the rainy season promoted shifts in community structure with an increased occurrence of taxa tolerant to low salinity conditions. Harmful and bloom-forming phytoplankton, including Thalassiosira sp. and Bellerochea sp., as well as potentially toxic dinoflagellates such as Prorocentrum spp. and Gymnodinium sp., were detected, particularly during the dry season. The proliferation of these taxa may contribute to oxygen depletion and increase the risk of toxin accumulation in shellfish. Therefore, continuous monitoring of water quality and phytoplankton communities, particularly during the dry season, combined with effective management of nutrient inputs and environmental conditions, is essential to reduce ecological risks, ensure food safety, and support the long-term sustainability of shellfish aquaculture.

Metal- and Solvent-Free Synthesis of Tris(indolyl)methanes via the Reaction of Indoles with Trimethyl Orthoacetate

Piyada Subgerd , Uthaiwan Sirion, Jaray Jaratjaroonphong — Mon, 11 May 2026 00:00:00 +0700

Background and Objectives: Tris(indolyl)methanes (TIMs) are important heterocyclic scaffolds widely found in biologically active molecules and functional organic materials. Indole-containing compounds are prevalent in natural products, pharmaceuticals, and agrochemicals, and exhibit diverse biological activities, including anticancer, antimicrobial, antiviral, and anti-inflammatory effects. In particular, TIM derivatives have attracted considerable attention due to their pharmacological potential and their utility as key intermediates in organic synthesis, as well as their applications in materials science such as organic electronics and molecular sensors. TIMs are typically synthesized via acid-catalyzed Friedel–Crafts reactions between indoles and electrophilic carbon sources such as orthoesters. However, many reported methods rely on strong mineral acids, expensive metal catalysts, elevated temperatures, or large amounts of organic solvents, which pose environmental concerns and operational limitations. Although recent advances in green chemistry have promoted solvent-free and metal-free approaches, the development of simple, efficient, and environmentally benign methods for TIM synthesis remains a significant challenge. Therefore, this study aims to develop a straightforward and efficient method for the synthesis of tris(indolyl)methanes via the reaction of indoles with trimethyl orthoacetate under solvent-free and metal-free conditions, providing a practical and sustainable alternative with operational simplicity and improved environmental compatibility.

Methodology: The synthesis of tris(indolyl)methane derivatives was achieved through a Friedel–Crafts-type electrophilic substitution reaction between indoles and trimethyl orthoacetate. In this study, p-toluenesulfonic acid (p-TSA) was employed as an inexpensive and readily available organic acid catalyst to promote the reaction. The reactions were carried out under solvent-free conditions at room temperature in order to enhance the environmental compatibility of the process and to simplify the experimental procedure. Typically, indole (1 mmol) and trimethyl orthoacetate (1 mmol) were mixed in the presence of 30 mol% p-toluenesulfonic acid, and the reaction progress was monitored over time. To evaluate the generality and applicability of the developed protocol, a variety of indole substrates bearing different substituents on the aromatic ring were examined. The corresponding tris(indolyl)methane products were isolated and characterized using spectroscopic techniques.

Main Results: The developed protocol successfully afforded the desired tris(indolyl)methane derivatives within relatively short reaction times ranging from 30 minutes to 2 hours. The products were obtained in moderate to good yields, typically ranging from 40% to 80%, depending on the electronic nature of the substituents on the indole ring. Electron-donating and electron-withdrawing substituents were generally tolerated under the optimized conditions, demonstrating the versatility of the reaction. Importantly, the transformation proceeded efficiently under mild conditions without the need for metal catalysts, external solvents, or elevated temperatures. The solvent-free reaction environment also contributed to a simpler reaction setup and reduced waste generation. These results indicate that trimethyl orthoacetate serves as an effective electrophilic partner for the formation of tris(indolyl)methane derivatives through a Friedel–Crafts-type process.

Conclusions : This study describes a convenient and environmentally friendly method for the synthesis of tris(indolyl)methane derivatives through the reaction of indoles with trimethyl orthoacetate. The protocol employs p-toluenesulfonic acid as an inexpensive catalyst and operates efficiently under metal-free and solvent-free conditions at room temperature. The method provides moderate to good yields and exhibits a reasonable substrate scope with various substituted indoles. Compared with previously reported procedures, the present approach offers several advantages, including operational simplicity, mild reaction conditions, reduced use of hazardous reagents, and improved environmental compatibility. Therefore, this method represents a practical and sustainable alternative for the preparation of tris(indolyl)methane derivatives and may be useful for future applications in medicinal chemistry and organic synthesis.

Urban PM2.5 Concentration Estimation Based on Remote Sensing and Land Use Regression Model: Evidence from the Bangkok Metropolitan Region

Thanyarat Chaiyakarm, Pannee Cheewinsiriwat , Ekkamol Vannametee — Fri, 15 May 2026 00:00:00 +0700

Background and Objectives: Fine particulate matter (PM_2.5) pollution has been a persistent environmental issue in Thailand for over two decades. In 2023, Thailand ranked fifth in Southeast Asia among countries experiencing the worst air quality for over 30 consecutive days. Bangkok and its metropolitan area, as the country’s economic, political, social, and cultural hub, have undergone rapid urban expansion, intensifying cumulative pollution and continuously affecting residents’ quality of life. In 2022, only three months recorded good air quality. The highest PM_2.5concentrations occurred during March and April, particularly in roadside areas, where levels began to pose adverse health impacts and showed an increasing trend. However, the limited number of air quality monitoring stations restricts comprehensive surveillance and accurate assessment of PM_2.5concentrations. Aerosol Optical Depth (AOD) data obtained from the Himawari-8 satellite are widely recognized as a valuable alternative for estimating PM_2.5levels in areas without ground-based monitoring due to their broad spatial coverage. Meteorological conditions and urban factors also influence PM_2.5concentrations. Land Use Regression (LUR), extensively applied in previous studies, has demonstrated strong capability in explaining PM_2.5factors. Therefore, this study aims to examine the relationships among PM_2.5concentrations, AOD data, meteorological factors, and urban factors in the Bangkok Metropolitan Region. The LUR model is developed to estimate PM_2.5levels in unmonitored areas and to generate monthly concentration maps covering the entire study area.

Methodology: Data were collected from 73 air quality monitoring stations operated by the Pollution Control Department and the Bangkok Metropolitan Administration to develop a PM_2.5database. Meteorological factors—including wind speed, temperature, relative humidity, and air pressure—were compiled. In addition, Aerosol Optical Depth (AOD) data obtained from Himawari-8 in NetCDF format, with a spatial resolution of 5 × 5 square kilometers, were extracted at station locations. Subsequently, all datasets were organized into daily and monthly records. Urban factors comprised traffic speed, population density, and land use. For spatial analysis, buffer zones with radius of 200, 400, 600, 800, and 1,000 meters were mapped around each station. The buffer distance showing the highest correlation with PM_2.5for each factor was then selected for model development. To investigate factor associations, Pearson’s correlation coefficient at a 0.01 significance level was applied to identify variables significantly associated with PM_2.5. These selected predictors were then incorporated into a Land Use Regression (LUR) model using stepwise selection to construct a multiple linear regression equation for estimating PM_2.5concentrations. Model performance was subsequently evaluated using R² and Root Mean Square Error (RMSE), together with 10-fold and leave-one-out cross-validation (LOOCV). Finally, the resulting regression equation was used to produce monthly PM_2.5maps for the study area.

Main Results: The results of PM_2.5concentration estimation in the Bangkok Metropolitan Region area during 2019–2023 reveal pronounced spatial and temporal variability, with peak concentrations consistently occurring in the winter season each year. Analysis of the relationships between PM_2.5and meteorological factors indicates that air pressure and relative humidity were the most influential factors. Relative humidity exhibited a significant negative correlation with PM_2.5concentrations, whereas air pressure showed a significant positive correlation. Regarding urban factors, several land use categories were significantly associated with PM_2.5, reflecting the role of human activities in shaping pollution distribution patterns. Traffic speed and population density demonstrated weak negative correlations; Aerosol Optical Depth (AOD) data derived from Himawari-8 showed positive correlations with PM_2.5at all monitoring stations. Although the strength of these correlations was moderate, their temporal consistency highlights the potential of AOD as a surrogate for ground-based monitoring data in areas with limited station coverage. The Land Use Regression (LUR) model yielded an R² value of 0.339 for the combined five-year dataset, with a Root Mean Square Error (RMSE) of 9.971. The highest R² value (0.456) was observed in 2020, demonstrating the model’s capacity to explain PM_2.5variability in a complex urban environment. The model outputs were further applied to develop monthly PM_2.5maps, providing a foundational dataset for area-based air quality management planning.

Conclusions: The research findings indicate that PM_2.5concentrations exhibit pronounced seasonal variability, particularly during the winter period. Correlation analysis revealed that meteorological factors are significant determinants, exerting statistically significant influences on PM_2.5levels. In contrast, urban factors play a comparatively lesser role in explaining pollution variability. Aerosol Optical Depth (AOD) data derived from the Himawari-8 demonstrated consistent correlations with PM_2.5concentrations across all monitoring stations over multiple time periods. The Land Use Regression (LUR) model achieved a moderate level of predictive performance and was capable of generating monthly PM_2.5concentration maps. These outputs can effectively support localized air quality monitoring and management strategies.

A Comparison of Efficiency for Homogeneity of Variance Tests under Mixture Normal Distribution

Vanida Pongsakchat, Natthawat Wanthong — Fri, 15 May 2026 00:00:00 +0700

Background and Objectives: In statistical data analysis, the validity and reliability of conclusions derived from hypothesis testing depend on several crucial assumptions. One of the most important assumptions is the Homogeneity of Variance, which requires that the variance of the data in every studied population must be equal. This assumption is a necessary condition for widely used statistical tests, such as the t-test for comparing the means of two populations and F-test for comparing the means of three or more populations. Statisticians have developed several testing methods to verify this assumption. Traditional methods like Bartlett’s test are well-known for having high efficiency when the data follow a normal distribution but are extremely sensitive to violations of normality. Consequently, other tests with greater robustness have been proposed, such as Levene’s test (in both mean-based and median-based forms), the Non-parametric Levene’s test, O'Brien’s test, and the Fligner-Killeen test, each of which has different characteristics and efficiency depending on the nature of the data, such as skewness, kurtosis, and sample size.Currently, real-world data are often more complex than a single probability distribution, frequently appearing in the form of a mixture distribution. The mixture normal distribution is a very common probability distribution found in empirical research, such as in financial markets, medicine, and epidemiology. Using standard normal models is therefore insufficient for explaining these data. Studying the efficiency of homogeneity of variance tests under mixture distributions is critically necessary. This research, therefore, aims to study and compare the efficiency of six homogeneity of variance tests: Bartlett’s test, Mean-based Levene’s test, Median-based Levene’s test, Non-parametric Levene’s test, O'Brien’s test, and the Fligner-Killeen test, focusing on three population groups under a two-component mixture normal distribution, and considering the Type I error probability and statistical power criteria across various simulated situations.

Methodology: Data for the three groups were generated from a two-component mixture normal distribution with mixing proportions (p ) were p=0.5 and p=0.8. The mean for all three groups was set to 8.8 with initial variances specified at 5, 10, and 20. Equal sample sizes were assigned to every group as (10,10,10), (20,20,20), (40,40,40), (60,60,60), (80,80,80), and (100,100,100). The variance ratios were set at 1:1:1, 1:1:2, 1:1:4 and 1:2:4. To consider the ability to control Type I error, Bradley’s liberal criterion was used at a significance level of 0.05. A test considered capable of controlling type I error must have an estimated type I error probability within the range of [0.025, 0.075]. In considering statistical power, only tests that passed Bradley’s criterion were considered; an efficient test must have an estimated statistical power of not less than 0.80, and the method with the highest value is considered the most efficient.

Main Results: The study demonstrates that the efficiency of all six testing methods differs depending on the mixing proportion and sample size. When considering the ability to control type I error: in the case where p=0.8 , the Median-based Levene’s test, Non-parametric Levene’s test, O’Brien’s test, and Fligner-Killeen test were robust, being able to control Type I error across all sample sizes, while Bartlett’s test required a sample size of 60 or more to maintain control. The Mean-based Levene’s test could not control the error in any situation. In the case where p=0.5 , the Median-based Levene’s test, Non-parametric Levene’s test, and O’Brien’s test could control Type I error across all sample sizes. The Mean-based Levene’s test began to maintain control when . However, Bartlett’s test and the Fligner-Killeen test could not control the error under this mixing proportion. In considering the estimated statistical power, only tests that could control Type I error were examined. It was found that the estimated statistical power increases according to sample size and the difference in variances. At p=0.8 : All four methods capable of controlling Type I error required a sample size of for the 1:1:2 ratio and for the 1:1:4 ratio to achieve statistical power higher than 0.8. Bartlett’s test provided the highest statistical power in conditions where it could control Type I error. At p=0.5, O’Brien’s test was the most prominent, providing statistical power higher than 0.8 when at a 1:1:2 ratio and at a 1:1:4 ratio, which outperformed all forms of Levene’s test.

Conclusions : The results of this study: 1. Bartlett’s Test: Although highly efficient in some conditions ( p=0.8 and large sample size), it cannot control Type I error when the data have a mixture distribution with equal mixing proportions, making it a high risk for application with complex data. 2. Mean-based Levene’s Test: Highly sensitive to mixing proportions and only applicable when the mixture distribution is balanced and the sample size is moderate or larger. 3. Fligner-Killeen Test: Performs well only in cases where the mixing proportion p=0.8 . 4. O’Brien’s Test: This method is identified as the most suitable and efficient overall. It not only successfully controlled the Type I error under all studied conditions but also provided the highest statistical power in nearly every scenario, particularly with small to moderate sample sizes.

Enhancement of Phosphate Adsorption Efficiency in Synthetic Wastewater Using Water Treatment Plant Sludge Combined with Cocoa Husk

Supawadee Noinumsai, Sitthisak Waipib , Chatchai Chada — Mon, 18 May 2026 00:00:00 +0700

Background and Objectives: Cocoa husks are agricultural waste with potential for use as phosphate adsorbents in wastewater treatment. It can be applied as adsorbent material for pollutants and nutrients in wastewater, making it an alternative to the development of natural adsorbent materials that are low-cost and non-toxic to the environment. Currently, water pollution problems, especially phosphate nutrient contamination, have intensified progressively. Prior research has demonstrated that cocoa husks can adsorb phosphate from water, and surface modification with chitosan significantly improves adsorption efficacy. Additionally, sludge from water treatment plants contain mineral compositions suitable for phosphate binding. Therefore, this study aimed to investigate the enhancement of phosphate adsorption efficiency in synthetic wastewater using chitosan-coated adsorbent materials made from sludge from water treatment plants (S) combined with cocoa husks (C).

Methodology: The methodological procedures involved testing the molding of adsorbent materials between soil sediment from water treatment plant sludge (S) and cocoa husks (C), both without chitosan surface modification and with chitosan surface modification at ratios of 100:0, 70:30, 50:50, 30:70, and 0:100 by weight. The materials were then molded by firing at temperature of 500°C for 2 hours. The study examined equilibrium time of adsorption at various time intervals, including 1, 2, 4, 8, 12, 16, 20, 24, 28, 32, 36, 40, 44, and 48 hours respectively.To determine the optimal contact time and adsorption isotherms of phosphate in synthetic wastewater using Langmuir and Freundlich models. Residual phosphate concentration was analyzed using a spectrophotometer at a wavelength of 880 nanometers, followed by calculation of the phosphate removal percentage in wastewater.

Main Results: The molding study of adsorbent materials showed that adsorbent materials at ratios S-C (100:0) and S-C (70:30) could be molded into pellets with stability and water insolubility, making them suitable for selection in subsequent procedures. Meanwhile, ratios S-C (50:50) and S-C (30:70) could not maintain form and dissolved in distilled water. This dissolution behavior was attributed to the higher organic content in these ratios, which compromised the structural integrity and thermal stability of the pellets during the firing process. All adsorption experiments were conducted at an initial pH of 5, which was selected based on preliminary optimization studies as the optimal condition for maximum phosphate solution removal efficiency. Under these conditions, the protonation of chitosan's amino groups was maximized, facilitating strong electrostatic interactions with phosphate anions. For the equilibrium time study of uncoated and chitosan-coated adsorbent materials revealed that coated adsorbent materials at ratio S-C(70:30)+C and S-C(100:0) reached equilibrium faster than uncoated adsorbent materials, achieving equilibrium within 16 hours and 24 hours respectively. This occurred because surface modification with chitosan resulted in faster equilibrium time, indicating adsorption kinetics. The adsorption performance study found that adsorbent materials with chitosan solution surface modification could adsorb phosphate concentration in synthetic wastewater higher than adsorbent materials without chitosan coating, with phosphate removal efficiencies of 60.45% and 47.68%, respectively. This demonstrated that surface modification with chitosan coating increased adsorption performance by 27.00% compared to the original. Additionally, analysis of the adsorption isotherm equations demonstrated the adsorption mechanisms of both chitosan-coated and uncoated adsorbent materials. Based on R² values closest to unity, the S-C (70:30)+C ratio showed the best fit, with the adsorption process conforming more closely to the Langmuir isotherm model (R² = 0.9950) than to the Freundlich isotherm (R² = 0.9628). The maximum adsorption capacity (q_max) value is 72.4638 mg/g, indicating that the adsorption on the surface of the adsorbent material is characterized by a single layer of adsorption.

Conclusions: Surface modification with chitosan coating enhanced phosphate adsorption performance due to the protonation of amino groups (NH₂) under slightly acidic conditions, enabling strong electrostatic attraction with negatively charged phosphate ions. This material demonstrates high potential for application as a low-cost adsorbent in wastewater treatment, while simultaneously adding value to agricultural waste and reducing local environmental problems.

Biology and Culture of the Fall Armyworm 𝙎𝙥𝙤𝙙𝙤𝙥𝙩𝙚𝙧𝙖 𝙛𝙧𝙪𝙜𝙞𝙥𝙚𝙧𝙙𝙖 with Mung Bean-Based Artificial Diet in Laboratory Conditions

Kasem Kongnirundonsuk, Nakorn Pradit — Wed, 20 May 2026 00:00:00 +0700

Background and Objectives: The fall armyworm (FAW), Spodoptera frugiperda, is an important pest that poses a significant threat to maize and other economic crops, including rice, sugarcane, cotton, and peanuts, due to its polyphagous feeding habits and widespread global distribution, including Thailand. Currently, global research focuses on management of this pest through chemical and biological control methods, including the use of natural enemies such as nucleopolyhedrovirus (NPV) and Bt (Bacillus thuringiensis), as well as on studies of its nutritional requirements for growth and behavior under natural and laboratory conditions. Such studies require large numbers of test insects with controlled variability in age and health to ensure accurate experimental results. Thus, artificial diets play a critical role in laboratory rearing systems by providing standardized and consistent nutritional composition, eliminating the variability associated with natural host plants, and enabling mass production of insects throughout the year. At present, artificial diets for S. frugiperda commonly utilize protein sources derived from various legumes. However, some of these legumes are expensive and not readily available in Thailand, limiting their practical application. Therefore, this study aimed to evaluate the effectiveness of a mung bean (Vigna radiata (L.) R. Wilczek) -based artificial diet on the growth and reproduction of S. frugiperda under laboratory conditions in order to identify a cost-effective alternative protein source with high protein content, low cost, and wide availability in Thailand.

Methodology: The biological development and reproductive performance of S. frugiperda reared on a mung bean–based artificial diet were investigated under laboratory conditions at 25±2 °C and 80–90% relative humidity. One-day-old eggs (n = 100) were collected and observed under a stereomicroscope to determine the egg incubation period. After hatching, individual larvae were transferred to 59.14 ml transparent plastic cups, with one larva per cup to prevent cannibalism. Larvae were fed a mung bean–based artificial diet, which was replaced every three days until the larvae reached the pupal stage. Developmental duration and mortality were recorded daily. Upon pupation, pupal width and length were measured to compare the sizes of male and female pupae.. During the adult stage, newly emerged male and female moths were paired, and a total of 20 mating pairs were placed in transparent plastic boxes (13.5 × 20.5 × 7 cm), with one pair per box, to determine the number of eggs laid by each female. Daily observations were conducted until adult mortality. All Data collected included developmental duration and mortality rate of each life stage of S. frugiperda, pupal size, adult longevity, and daily fecundity. Eggs laid on wax paper and plastic containers were counted daily to determine reproductive output. Fertility table parameters, including the net reproductive rate (R₀), intrinsic rate of increase (r_c), finite rate of increase (), and mean generation time (T_c), were calculated using demographic methods. Differences in pupal size and adult longevity between males and females were analyzed using Student’s t-test.

Main Results: The results indicated that S. frugiperda successfully completed its life cycle and exhibited rapid population growth when reared on the mung bean–based artificial diet under laboratory conditions. The life cycle consisted of four stages: egg, larva, pupa, and adult. The mean developmental duration from egg to adult emergence was 37.22±0.16 days. Fertility table analysis revealed a high reproductive potential of S. frugiperda reared on the mung bean diet. The net reproductive rate (R₀) was 384.86 offspring per individual female. The intrinsic rate of increase (r_c) was 0.13 day^-1, while the finite rate of increase () was 1.14 day^-1, indicating a rapid daily population growth rate under laboratory conditions. The generation time (T_c) was 45.77 days. Comparisons of pupal width and length between males and females revealed no statistically significant differences, whereas the adult lifespan of males was significantly shorter than that of females (p < 0.05).

Conclusions: Rearing S. frugiperda on a mung bean–based artificial diet effectively supported its growth, development, and reproduction under laboratory conditions. Although the developmental period of insects reared on the mung bean diet was longer than that reported for artificial diets based on white beans and pinto beans, the overall reproductive yield was higher. In addition, mung bean is inexpensive, widely available, and easily accessible in Thailand, making it a practical and economical protein source for artificial diet formulation. Therefore, the mung bean–based artificial diet represents a suitable and cost-effective alternative for the mass rearing of S. frugiperda in laboratory research and reduces reliance on expensive or imported protein sources.

A Hybrid Safeguarded Quasi-Newton Method for Nonlinear Systems with Engineering Applications

Apichat Neamvonk — Wed, 20 May 2026 00:00:00 +0700

Background and Objectives: Nonlinear systems of equations arise extensively in engineering, applied physics, computational science, economics, and industrial process modeling. Accurate and efficient numerical techniques for solving such systems are therefore essential in many scientific and engineering applications. Classical Newton’s method remains one of the most widely used approaches because of its rapid local quadratic convergence. However, its implementation requires repeated evaluations and factorizations of the Jacobian matrix at every iteration, which can become computationally expensive for large-scale or moderately complex nonlinear systems. In addition, Newton’s method is highly sensitive to the choice of initial approximations and may fail to converge when the Jacobian matrix is singular, ill-conditioned, or poorly approximated near the solution. To reduce computational cost, quasi-Newton methods such as Broyden’s method replace exact Jacobian evaluations with low-rank matrix updates. These methods significantly improve efficiency by avoiding repeated Jacobian computations, making them attractive for practical applications involving expensive function evaluations. Nevertheless, classical Broyden-type methods may experience instability, stagnation, inaccurate search directions, and error accumulation due to progressive deterioration of the inverse Jacobian approximation during the iterative process. Such difficulties often reduce convergence reliability, particularly for strongly nonlinear or ill-conditioned problems. Motivated by these limitations, this study proposes a Hybrid Safeguarded Quasi-Newton Method (HSQNM) that combines the efficiency of quasi-Newton updates with additional stabilization and safeguarding mechanisms to improve robustness, reliability, and overall convergence performance.

Methodology: The proposed HSQNM integrates an inverse Broyden rank-one update with three complementary stabilization components: adaptive damping, curvature-based safeguarding, and restart-based recalibration. The adaptive damping mechanism dynamically adjusts the step length according to the behavior of the nonlinear residual, ensuring sufficient residual reduction and preventing excessively large iterative steps that may lead to divergence. A curvature-monitoring criterion is incorporated to detect unstable or unreliable updates when the secant condition becomes nearly degenerate or when the approximate inverse Jacobian loses numerical quality. Under such conditions, a safeguarding strategy modifies the update process to maintain stability and improve descent behavior. Furthermore, a restart mechanism is introduced to periodically reconstruct the inverse Jacobian approximation whenever deterioration or stagnation is detected. This recalibration process reduces accumulated numerical errors and restores the quality of the iterative search direction. Theoretical analysis of the proposed method is developed under standard assumptions of Lipschitz continuity and nonsingularity of the Jacobian matrix in a neighborhood of the exact solution. Local convergence properties are established to demonstrate that the proposed hybrid strategy preserves the desirable convergence characteristics of quasi-Newton methods while improving robustness in challenging nonlinear regimes.

Main Results: To evaluate the effectiveness of the proposed algorithm, numerical experiments were conducted on a collection of benchmark nonlinear systems together with several engineering applications involving nonlinear mathematical models. The benchmark problems include systems with varying dimensions, nonlinearities, and conditioning characteristics to examine both efficiency and stability. Performance comparisons were carried out against the classical Broyden method and Newton’s method using metrics such as iteration count, residual norm reduction, Jacobian evaluations, computational cost, and convergence success rate. The numerical results demonstrate that HSQNM consistently achieves lower failure rates and improved convergence stability compared with the classical Broyden method. In many test problems, the proposed approach required fewer Jacobian recalculations while maintaining accurate convergence behavior. The adaptive damping strategy was particularly effective in preventing divergence for poor initial guesses, whereas the restart mechanism successfully mitigated stagnation caused by inaccurate inverse Jacobian approximations. Additional experiments on engineering nonlinear systems, including diode circuit analysis and radiative heat transfer models, further confirmed the practical applicability of the proposed method. In these applications, HSQNM achieved convergence accuracy comparable to Newton’s method but with reduced computational expense and improved robustness. The results also indicate that the safeguarding framework enhances reliability for ill-conditioned systems where traditional quasi-Newton methods may fail or exhibit unstable iterative behavior.

Conclusions: The proposed Hybrid Safeguarded Quasi-Newton Method provides a reliable and computationally efficient framework for solving moderate-scale nonlinear systems arising in engineering and scientific applications. By integrating adaptive damping, curvature-based safeguarding, and restart-based recalibration into the inverse Broyden update process, the method improves convergence robustness without sacrificing the computational advantages of quasi-Newton techniques. Numerical experiments demonstrate that the proposed approach offers better stability, lower failure rates, and competitive computational performance compared with classical Broyden-type methods. Moreover, the method retains convergence accuracy comparable to Newton’s method while reducing the frequency of expensive Jacobian evaluations. The flexibility of the hybrid safeguarding framework also suggests potential extensions to large-scale sparse nonlinear systems, optimization-based nonlinear solvers, and parallel scientific computing applications. Consequently, HSQNM represents a promising alternative for practical nonlinear system solving in modern computational engineering and applied mathematics.

A Better Lower Bound for the Number of Magic Squares of Order 4𝒌

Benjawan Tiamkaew, Boonyisa Saethao , Boorapa Singha — Wed, 20 May 2026 00:00:00 +0700

Background and Objectives: Magic squares of order n are n x n arrays filled with distinct positive integers 1,2,3,..., arranged in such a way that the sums of the integers in each row, column and each of the main diagonals are the same. This number is called the magic constant or magic sum which is equal to . Magic squares have a long history, the earliest recorded information about magic squares comes from China around 2200 BC and is called "Lo Shu". Magic squares were first mentioned in the western world in the works of Theon of Smyrna and were also used to help calculate horoscopes by a 9th-century Arab astrologer. Next, the work of Greek mathematician Moschopoulos in 1300 helped popularize knowledge about magic squares. Today, over 700 years later, teachers still use this knowledge in the classroom to solve problems and practice addition. The mathematical study of magic squares generally involves their construction, classification and enumeration. Although completely general methods for creating all the magic squares of all orders do not exist, there are studies on algorithms that can be used to construct magic squares in some orders. For example, the De la Loubère method (or Siamese method) is a simple algorithm for creating magic squares of any odd order by placing numbers sequentially, moving diagonally up and to the right, wrapping around the grid, and dropping down one cell when hitting a filled cell or boundary, this technique was observed by the French diplomat Simon de la Loubère in Siam (Thailand) in the late 17th century. The Strachey method constructs magic squares of singly even order n = 4m+2 by combining four smaller identical magic squares of order 2m+1 together to form one magic square of order 4m+2 . Then 0,

is added to each square and finally certain squares are swapped from the top subsquare to the bottom subsquare. Let M (2k, k) represent the set of 2k x 2k matrices in which every element is either 0 or 1, with the sums of each row and column equal to k and let be the set of magic squares of order 4k. In 2022, Oboudi showed that each matrix can be used to construct a magic square in . The author also showed that the number of magic squares of order 4k is at least . To prove this result, Oboudi demonstrated a comparison of the cardinality of three sets by showing that ., where H represents the set of some bipartite graphs that are defined in a specific way. However, comparing the cardinality of the three sets above causes the obtained lower bound to possibly deviate from the true value of . . To reduce this error, in this paper we use a new procedure that is simpler than Oboudi's technique and can obtain a lower bound of . that is closer to the true value.

Methodology: From Oboudi's inequality , in this study we will only consider the condition and aim at counting the number of matrices in . This procedure will not compare the number of elements in the set with other sets. Therefore, it reduces the potential for error. We will begin by considering the condition again. Some details of such a condition that have not been discussed in Oboudi's work will be presented in this study. We will present two algorithms for constructing matrices in and prove their correctness. Then we will show that the output matrices from both algorithms will always be different. After that, we will present formulas to count the number of matrices obtained from both algorithms, one of which is the lower bound that Oboudi has presented in 2022. Then we combine the number of matrices obtained from the two algorithms to find a new lower bound for the number of magic squares of order 4k.

Main Results: We have presented the conditions under which the inequality is true. After that, two algorithms for constructing matrices in are presented. Both algorithms produce different matrices. To obtain a new lower bound for the number of magic squares of order 4k , formulas for counting matrices obtained from both algorithms are investigated. The first is , which is the lower bound of Oboudi, and the second is . After combining the two formulas together, a modified lower bound for the number of magic squares of order 4k is introduced, which is

Conclusions: This research is a study of a lower bound for the number of magic squares of order 4k using the concept from Oboudi's work. The result obtained from this study is a lower bound that are closer to the true value of than the original values. The new lower bound is obtained by two algorithms of constructing matrices in M(2k,k). We found that the matrices obtained from the two algorithms are different and there are exactly matrices obtained from the first algorithm. This number is equal to the lower bound that Oboudi studied in 2022, but here we use a simpler proof. Moreover, we proved that the second algorithm can construct at least matrices. Finally, when the number of matrices obtained by the two constructions are combined, a new and better lower bound for the number of magic squares of order 4k is obtained.

Power 3 Mean Labeling of Some Special Graphs

Roongrat Samanmoo — Fri, 22 May 2026 00:00:00 +0700

Background and Objectives: Graph labeling, the assignment of integers to the vertices, edges, or both of a graph under specific conditions, has been a central concept in graph theory since its introduction in 1960. Mean labeling, a specific type of graph labeling, was introduced by Somasundaram and Ponraj in 2003 and has since been extensively explored for various graph classes. The concept of Power 3 mean labeling was recently introduced in 2020 by Sreeji and Sandhya. A function f is called a Power 3 mean Labeling of a graph G = (V, E) with p vertices and q edges if it is possible to label the vertices u V with distinct labels f(u) from 1,2, … , q+1 in such a way that when f(u) = x and f(v) = y, each edge e = uv is assigned a label defined by or .

or . In this case, f is a Power 3 mean labeling of G and G is called a Power 3 mean graph. Sreeji and Sandhya demonstrated that some graphs possess this property while others do not. Their subsequent work explored this labeling for the line graphs of certain graphs. Moreover, they considered Power 3 mean labeling for various graphs resulting from the duplication of graph elements. This paper contributes to this area of study by proving that Ladder graphs, Total graphs of a path and Chains of cycles are a family of Power 3 mean graphs. n–ladder graph can be defined as the Cartesian product of two paths, , and it is equivalent to the 2xn graph. Total graph of G, denoted by T(G), is the graph whose vertex set consists of all vertices and edges of G. Two vertices in T(G) are adjacent if and only if their corresponding elements in G are either adjacent or incident in G. Chain of m cycle of length n is a graph obtained by taking m copies of the cycle graph (n 3, m 1), arranging them in linear sequence, and identifying exactly one vertex of the i-th copy with one vertex of the (i+1)-th copy for i = 1, …, m-1. We denote this chain by .

Methodology: To study Power 3 mean graphs, we begin by considering the relationships between the numbers assigned to an edge uv when the endpoints are f(u) = i and f(v) = i+m. For m=1,…,5, the lower and upper bounds of the possible 3 mean values can be determined, and a labeling pattern for the vertices and edges that satisfies the given conditions can be identified. Next, for n 1, we prove by induction on n to show that and are Power 3 mean graphs. Finally, for n 3 and m 1 the Power 3 mean labeling of is considered in two cases: when n is even and when n is odd.

Main Results: We establish four lemmas describing the possible integer values assigned to an edge uv when f(u)=i and f(v) = i+m, the results are summarized as follows:

(1) m = 1 , f(uv) = i or i +1

(2) m = 2 or 3 , f(uv) = i +1 or i +2

(3) m = 4 , f(uv) = i +2 or i +3

(4) m = 5 , f(uv) = i +3 or i +4 when 3 i 9 and

, f(uv) = i +2 or i +3 when i 10.

We obtain three theorems presenting the Power 3 mean labeling functions of the ladder graph, the total graph of a path, and the chain of m cycles of length n.

Conclusions: The Ladder graph, Total graph of path and Chain of m cycle of length n are Power 3 mean graphs. To determine whether a class of graphs admits a Power 3 mean labeling, one common approach is to identify a suitable labeling pattern for the vertices and edges that satisfies the given conditions. This is typically achieved by analyzing the relationships among the integers assigned to the end vertices and the resulting values obtained for the corresponding edges. In the case of Power 3 mean labeling, when there is a large difference between the numbers of edges and vertices, the structure of the graph becomes more complex. As a result, it becomes difficult to determine an appropriate labeling pattern for the vertices. This difficulty also arises in graphs that form a closed loop, where finding a consistent labeling pattern under the given conditions is particularly challenging. The study of graph labeling remains a topic of significant interest among mathematicians, with research conducted under various conditions and on different types of graphs. For Power 3 mean graphs, there remain several classes of graphs that are of interest to investigate in order to determine whether they admit a Power 3 mean labeling.

Morphometric Comparison of Two Rare Catfishes, 𝑪𝒍𝒂𝒓𝒊𝒂𝒔 𝒏𝒊𝒆𝒖𝒉𝒐𝒇𝒊𝒊 and 𝑪𝒍𝒂𝒓𝒊𝒂𝒔 𝒎𝒆𝒍𝒂𝒅𝒆𝒓𝒎𝒂, from the Sirindhorn Swamp Forest, Narathiwat Province, Southern Thailand

Tue, 26 May 2026 00:00:00 +0700

Background and Objectives: Princess Sirindhorn Wildlife Sanctuary (Pru To Daeng) contains the largest remaining primary peat swamp forest in Thailand and represents a distinctive freshwater ecosystem characterized by acidic, waterlogged conditions and high organic matter. This habitat supports diverse freshwater fishes that are highly specialized for survival in hypoxic blackwater environments. Among them, walking catfishes of the genus Clarias exhibit notable morphological and physiological adaptations, including elongated bodies, scaleless skin, dorsoventrally flattened heads, four pairs of barbels, and suprabranchial organs that enable aerial respiration. The genus Clarias comprises approximately 62 species distributed across Africa and Asia. In Thailand, five species are recognized. While C. batrachus and C. macrocephalus are economically important, C. meladerma, C. leiacanthus, and C. nieuhofii are relatively rare and strongly associated with swamp forest habitats. Although C. nieuhofii and C. meladerma are classified as Least Concern, both exhibit restricted distributions and high ecological specialization, making them particularly sensitive to habitat alteration. Morphometric analysis provides a robust quantitative approach to investigating body-shape variation in relation to taxonomy, ecology, and adaptive differentiation. However, morphometric data for Clarias remain limited, as previous studies have focused mainly on commercially important species, leaving habitat-specialist taxa poorly documented. This study aims to (1) quantify and compare the morphometric characteristics of C. nieuhofii and C. meladerma, (2) examine sexual dimorphism within each species, and (3) evaluate the effectiveness of principal component analysis (PCA) in discriminating interspecific morphological variation.

Methodology: Twenty-six slender walking catfish (Clarias nieuhofii) and thirty-four black catfish (Clarias meladerma) were collected from the Sirindhorn Swamp Forest, Narathiwat Province, Thailand. External morphological traits, including coloration, body form, and fin and barbel characteristics, were documented. Twenty-four morphometric characters were measured using Vernier calipers following the Department of Fisheries (2007). These measurements included major body lengths, head dimensions, fin lengths, barbel lengths, and body depths, and were expressed as %SL and %HL. Mean and standard deviation were calculated, and differences between sexes and between species were analyzed using independent t-tests. To remove size effects prior to multivariate analysis, morphometric variables were adjusted using Turan’s (1999) ratio-based log transformation. Principal component analysis (PCA) was then applied to evaluate morphological separation between the two species.

Main Results: Clear interspecific differences in external morphology were observed between Clarias nieuhofii and C. meladerma. Clarias nieuhofii exhibited an elongate, slender body with a laterally flattened profile and distinct white neuromast markings arranged in 13–20 rows along the lateral surfaces. The head was relatively small and blunt, bearing four pairs of long barbels. The dorsal, anal, and caudal fins were continuous, with a rounded caudal fin tip, and body coloration ranged from dark to reddish-brown. In contrast, C. meladerma displayed a shorter, more robust body form with uniform dark black coloration and no visible spots. This species possessed serrated pectoral fin spines and a comparatively taller dorsal fin, resulting in a stockier appearance typical of non-slender walking catfishes. Sexual dimorphism in C. nieuhofii was present but limited. Of the 24 morphometric variables analyzed, four differed significantly between sexes. Males exhibited significantly greater total length (TL), standard length (SL), and lower jaw length (LLJ), indicating larger overall body size, whereas females showed significantly larger eye diameter (ED). All remaining variables, including head dimensions, body depth, fin lengths, and caudal peduncle depth, did not differ significantly. Principal component analysis (PCA) revealed partial separation between males and females, suggesting modest sex-related morphological variation. In contrast, sexual dimorphism in C. meladerma was minimal. Among all variables examined, only standard length (SL) differed significantly between sexes, with males showing slightly greater values. All other measurements, including head proportions, fin dimensions, eye diameter, and barbel lengths, showed no significant differences. PCA supported these results, as male and female individuals exhibited substantial overlap in multivariate space. Interspecific comparisons revealed pronounced morphological divergence between the two species. Twenty-one of the 24 morphometric variables differed significantly, with measurements related to head size, body depth, fin lengths, pelvic and anal regions, and barbel lengths consistently higher in C. nieuhofii, reflecting its elongated morphology. Only TL, caudal peduncle depth (CPD), and mouth width (WM) showed no significant differences. PCA further confirmed interspecific differentiation, with the first two principal components explaining 59.44% of the total variance, and PC1 clearly separating the species, despite minor overlap.

Conclusions: Morphometric comparison of Clarias nieuhofii and C. meladerma from the Sirindhorn Swamp Forest was based on 24 variables expressed as %SL and %HL. Sexual dimorphism was evident in C. nieuhofii, with significant differences in TL, SL, ED, and LLJ, whereas C. meladerma showed minimal sexual differentiation, with significance detected only in SL. Interspecific comparisons revealed significant differences in 21 of the 24 morphometric variables. Principal component analysis (PCA) explained 59.44% of the total variance and clearly separated the two species in multivariate space. These results indicate pronounced morphological divergence likely associated with ecological specialization in peat-swamp habitats and provide baseline data for future conservation and monitoring efforts.

Integrating ARIMA and GARCH Components within Recurrent Neural Cells for Enhanced Time-Series Prediction

Kunjira Kingphai — Thu, 28 May 2026 00:00:00 +0700

Background and Objectives: Financial time-series forecasting remains challenging due to the complex interaction between linear dependencies, nonlinear temporal patterns, and time-varying volatility in financial markets. Traditional statistical models such as Autoregressive Integrated Moving Average (ARIMA) and Generalized Autoregressive Conditional Heteroskedasticity (GARCH) provide interpretable representations of linear dynamics and volatility clustering but often struggle to capture nonlinear relationships in high-frequency data. Deep learning approaches, particularly Recurrent Neural Networks (RNNs) and Long Short-Term Memory (LSTM) models, can learn nonlinear sequence representations but typically rely on implicit feature learning without incorporating established statistical structures. Most hybrid approaches use pipeline-based integration, where statistical features are computed externally and fed into neural networks, limiting interaction between statistical structure and neural dynamics. This study develops a unified forecasting framework that integrates classical time-series operators directly within recurrent neural network cell dynamics, enabling joint modeling of temporal dependencies, nonlinear sequence dynamics, and time-varying uncertainty within a single end-to-end trainable system.

Methodology: This study proposes a statistically augmented recurrent forecasting framework in which classical statistical operators are embedded directly into the hidden-state transitions of recurrent neural network architectures. Six hybrid models are developed by integrating ARIMA and volatility-based components into both RNN and LSTM structures, resulting in ARIMA–RNN, Volatility–RNN, GARCH–RNN, ARIMA–LSTM, Volatility–LSTM, and GARCH–LSTM architectures. In these models, statistical coefficients are treated as trainable parameters and jointly optimized with neural network weights using backpropagation through time, enabling unified end-to-end learning. Empirical evaluation is conducted using intraday financial data sampled at five-minute intervals from seven highly liquid U.S. technology stocks commonly referred to as the "Magnificent Seven": Apple, Microsoft, Amazon, Alphabet, Meta, NVIDIA, and Tesla. Each observation includes open, high, low, close, and trading volume variables (OHLCV). Time-series forecasting is formulated using a sliding-window approach with a look-back window of 30 time steps. The dataset covers the period from January to September 2025 and is chronologically divided into training and testing sets using an 80/20 temporal split to avoid look-ahead bias. Time-series cross-validation with an expanding window strategy is used for model evaluation. Forecasting performance is assessed using Root Mean Squared Error (RMSE) and Mean Absolute Percentage Error (MAPE), and statistical significance is evaluated using the Scheirer-Ray-Hare test within each architecture group and Wilcoxon signed-rank tests between groups.

Main Results: The LSTM Baseline achieves the lowest overall average RMSE of 3.4955 and MAPE of 0.0080, outperforming all hybrid configurations in aggregate. Asset characteristics are the primary driver of forecasting accuracy, confirmed by a dominant stock effect across both architecture groups (Scheirer-Ray-Hare: H > 18, p < 0.005), while model selection does not produce statistically significant differences overall. LSTM-based models generally achieve lower average errors than ungated RNN architectures (RMSE 4.038 vs 4.233; MAPE 0.0092 vs 0.0107), though this difference is not statistically significant (Wilcoxon: W = 4890, p = 0.925). Statistically augmented models do not universally outperform baselines. Notably, GARCH–LSTM records the highest RMSE of 14.927 on META, suggesting that misalignment between statistical priors and event-driven volatility can amplify forecasting errors. However, during high-volatility periods, GARCH–LSTM achieved a lower MAE of 3.270 compared to the LSTM Baseline of 4.756, suggesting selective advantages for hybrid models under elevated market volatility.

Conclusions: The findings show that no single architecture consistently dominates across all assets, reflecting the heterogeneous nature of financial markets. The LSTM Baseline achieves the strongest overall performance, outperforming all hybrid configurations in aggregate. Statistical augmentation does not universally improve forecasting accuracy, and these results suggest that effective forecasting depends more on aligning model design with asset-specific characteristics than on increasing model complexity. However, hybrid models such as GARCH–LSTM showed selective advantages during high-volatility periods, suggesting that selective deployment during elevated market conditions may be more practical than general-purpose use. Several limitations should be noted. First, the evaluation is restricted to large-cap U.S. technology stocks at a single five-minute frequency, and it remains unclear whether the findings generalize to other asset classes or sampling intervals. Second, the embedded statistical coefficients are not explicitly constrained to satisfy classical parameter restrictions, as they are treated as learnable weights optimized through backpropagation rather than statistically inferred parameters. Future research should explore the application of this framework to broader asset classes such as bonds, commodities, and cryptocurrencies, as well as alternative sampling frequencies. Regime-aware model selection, where statistical priors are adaptively weighted based on real-time market conditions, represents a promising direction. Additionally, predicting returns rather than raw prices should be investigated to better align with the stationarity assumptions underlying classical ARIMA and GARCH formulations.