PERFORMANCE EVALUATION OF WATER QUALITY MONITORING SYSTEM IN NATURAL TROPICAL STREAMS

International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 218 PERFORMANCE EVALUATION OF WATER QUALITY MONITORING SYSTEM IN NATURAL TROPICAL STREAMS Emmanuel F. Akpan Department of Science Technology, Akwa Ibom State Polytechnic Ikot Osurua, Akwa Ibom State, Nigeria |  +234 708 458 0329| ✉ elderemmanuelak[email protected] Veronica M. Akpan Department of Science Technology, Akwa Ibom State Polytechnic Ikot Osurua, Akwa Ibom State, Nigeria |  +234 808 481 7644 |✉ veronicamkp[email protected]m Udeme U. Inyang Department of Science Technology, Akwa Ibom State Polytechnic Ikot Osurua, Akwa Ibom State, Nigeria | +234 806 748 0807|✉ [email protected] ARTICLE INFO ABSTRACT Paper ID: IJASTR68DF5F15D22E1 Received: 2025-09-06 Published: 2025-10-07 DOI: https://dx.doi.org/1 0.5281/zenodo.172842 28 Page No: 218-228 Obtaining dependable and economical water quality data continues to be a significant difficulty in resource-constrained areas, where traditional monitoring techniques are sometimes financially unfeasible and logistically intricate. This paper details the design, development, and field validation of an economical, microcontroller-driven electronic device for real-time monitoring of water quality in tropical surface waters, therefore solving this deficiency. The system, constructed with an Arduino Mega 2560, incorporates locally sourced sensors to assess electrical conductivity (EC), total dissolved solids (TDS), temperature, and turbidity, although pH testing is omitted due to sensor problems. Installed in four stream locations in Akwa Ibom State, Nigeria— Idim Atan, Inyang Udonwankwo, Idim Ikot Inyang, and Idim Nkap—from June 23 to June 26, 2025, the device captured 40 measurements (10 per location) over 8.5-hour intervals. Data analysis demonstrated an impeccable linear correlation between EC and TDS (R² = 1.000), confirming sensor accuracy, with EC values spanning from 600 to 820 µS/cm across locations, reaching a maximum at Idim Atan. A multiple regression model (R² = 0.513) demonstrated moderate prediction capability for electrical conductivity depending on turbidity and temperature. These findings highlight the device's potential as a reliable, accessible instrument for environmental evaluation in data-deficient tropical areas, providing insights into pollution dynamics. Keywords — Electrical conductivity, Low-cost sensors, Total dissolved solids, Tropical streams, Turbidity, Water pollution International Journal of Advanced Scientific and Technical Research Available online on http://www.rspublication.com/ijst/index.html ISSN 2249-9954 Cite This Paper : Emmanuel Friday Akpan, Veronica M. Akpan and Udeme U. Inyang (2025). "PERFORMANCE EVALUATION OF WATER QUALITY MONITORING SYSTEM IN NATURAL TROPICAL STREAMS". INTERNATIONAL JOURNAL OF ADVANCED SCIENTIFIC AND TECHNICAL RESEARCH (IJASTR), vol. 15, no. 5, 2025, pp. 218-228. DOI: https://dx.doi.org/10.5281/zenodo.17284228 International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 219 INTRODUCTION Access to clean and safe water constitutes a key worldwide challenge in the 21st century, as water pollution presents substantial challenges to human health, aquatic ecosystems, and sustainable development, especially in lowand middle-income nations. Accelerated urbanization, industrial effluents, agricultural runoff, and insufficient waste management have exacerbated the deterioration of natural water bodies, adding pollutants like suspended particles, fertilizers, heavy metals, and microbiological contaminants. In sub-Saharan Africa, where a significant share of rural populations relies on untreated surface water, the absence of dependable monitoring intensifies these challenges, rendering people susceptible to waterborne infections and ecological disruptions (Curtis et al. 2021). Tropical regions, such as Nigeria, encounter heightened challenges from climate unpredictability, seasonal flooding, and elevated humidity, which exacerbate pollutant dispersion and complicate water quality management (Mohammadpour et al. 2024). This circumstance necessitated the surveillance of water quality. Conventional water quality monitoring methods depend on laboratory techniques, including titration, gravimetric analysis, and spectrophotometry, which require costly equipment, specialized personnel, and intricate logistics for sample collection and analysis. These methodologies are frequently unfeasible in resource-constrained environments, resulting in a significant data deficit that obstructs prompt identification of contamination and informed policy formulation (Demetillo et al. 2019). The interval between sampling and the availability of results further constrains the capacity to address rising pollution issues, especially in distant regions where constant monitoring is essential (Okoye & Onyewuchi, 2017). This issue is particularly evident in rural Nigeria, where tropical streams are increasingly affected by agricultural runoff and industrial discharges, yet lack accessible monitoring facilities. As a result, computerized water quality monitoring was implemented to address the limitations associated with older methodologies. In this context, experts have suggested multiple digital methodologies for water monitoring. Notable among these unique concepts are the utilization of temperature, pH, and turbidity sensors, drone-based digital twin systems, and Internet of Things (IoT) systems (Gassava & Soussi, 2019). Recently, hybrid human-machine calorimetric techniques have been developed for water quality monitoring. These technologies enable communities, local governments, and non-experts to access real-time data, promoting participatory environmental management (Murti et al. 2024). This research aims to design, develop, and verify an economical electronic device for monitoring water quality. The objectives of this study are to utilize the device to measure multiple parameters (including EC, TDS, temperature, and turbidity) of water and to demonstrate its effectiveness across four Nigerian streams, thereby illustrating its reliability and efficacy in highly variable tropical stream environments. To tackle issues of affordability and reliability in current digital monitoring systems, the device will utilize locally-sourced materials, and its design will incorporate a range of commercially-available sensors interfaced with an Arduino microcontroller as the central component to enable continuous and automated data acquisition. This study is motivated by the fact that devices utilizing microcontrollers, such as Arduino and Raspberry Pi, can measure essential water parameters at a significantly lower cost than conventional instruments (Luyhimbi et al. 2022). This research will contribute International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 220 to the attainment of the United Nations Sustainable Development Goals (SDGs), specifically SDG 6 (Clean Water and Sanitation), which advocates for enhanced water quality via pollution mitigation, and SDG 9 (Industry, Innovation, and Infrastructure), which underscores sustainable technological advancements. The study will contribute to evidence-based water management by offering an accessible and reproducible monitoring tool, thereby empowering local stakeholders and supporting global initiatives to improve environmental sustainability and public health in data-deficient areas. Certain researchers (Gleick & Palaniappan, 2010) conducted a study to assess the physicochemical and bacteriological quality of drinking water provided by the municipality, from source to point of use, in Thulamela municipality, Limpopo Province, South Africa. The researchers evaluated community behaviors related to water collection and storage and identified the health risks to humans linked with water intake. Water quality assessment was conducted on 114 samples. Questionnaires were employed to ascertain the community's practices regarding water transportation from the source to the point of usage and storage activities. A significant number of homes indicated persistent interruptions in water delivery, and the majority (92.2%) did not purify their water prior to consumption. Although E. coli and total coliform were absent in the water samples from the source (dam), a majority of samples from street taps and domestic storage containers exhibited significant contamination with elevated levels of E. coli and total coliform. The concentrations of E. coli and total coliform observed in the wet season exceeded those recorded in the dry season. The concentrations of trace metals in the drinking water samples were within the acceptable limits established by both the South African National Standards and the World Health Organization. The assessed noncarcinogenic effects, determined via hazard quotient toxicity potential and cumulative hazard index for drinking water through ingestion and cutaneous exposure, were below one, indicating that water consumption presents no substantial non-carcinogenic health risk. Intermittent disruptions in municipal water supply and specific water transportation and storage techniques by community members heightened the potential of water pollution. The researchers advocate for a more reliable provision of treated municipal water in Limpopo province and the education of communities on hygienic methods for transporting and storing drinking water from the source to the point of use. Researchers (Junk et al. 2010) employed machine learning (ML) methodologies to assess drinking water quality and pinpoint pollution sources adjacent to a chromite mine in Iran. Human health hazards were evaluated utilizing both deterministic and probabilistic methodologies. The findings indicated that the quantities of calcium (Ca), chromium (Cr), lithium (Li), magnesium (Mg), and sodium (Na) in the water samples beyond international safety guidelines. The Unweighted Root Mean Square Water Quality Index (RMS-WQI) and the Weighted Quadratic Mean Water Quality Index (WQM-WQI) classified all water samples as 'Fair', yielding average scores of 67.95 and 67.19, respectively. Among the evaluated machine learning models, the Extra Trees (ET) method was identified as the foremost predictor of Water Quality Index (WQI), with magnesium (Mg) and strontium (Sr) as significant variables affecting the outcomes. Principal component analysis (PCA) revealed three separate clusters of water quality metrics, underscoring the impacts of both local geology and human International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 221 activities. The maximum average hazard quotient (HQ) for chromium (Cr) was 1.71 for children, 1.27 for adolescents, and 1.05 for adults. The Monte Carlo simulation for health risk assessment revealed median hazard indices (HI) of 4.48 for children, 3.58 for teenagers, and 2.98 for adults, all over the tolerable threshold of 1. The total carcinogenic risk (TCR) surpassed the EPA's tolerable threshold for 99.38% of children, 98.24% of teenagers, and 100% of adults, with arsenic (As) and chromium (Cr) identified as the primary culprits. EXPERIMENTAL PROCEDURE DEVICE DESIGN AND ARCHITECTURE The water quality monitoring system was designed as a compact, cost-effective, multiparameter sensing platform optimized for use in natural tropical freshwater habitats. The system architecture incorporates a collection of commercially accessible sensors connected to a programmable microcontroller to provide continuous, automatic data capture. The primary element is an Arduino Mega 2560 microcontroller. This microcontroller was chosen for its 54 digital input/output pins, 16 analog inputs, low power consumption (about 50 mA at 5V), and open-source programmability, which guarantees flexibility and scalability. The sensing array consists of the following components: an analog electrical conductivity (EC) sensor with builtin signal conditioning, an analog total dissolved solids (TDS) sensor calculated from EC measurements using a calibrated conversion factor, a waterproof DS18B20 digital temperature sensor utilizing the 1-Wire protocol with a resolution of ±0.5°C, and an infrared turbidity sensor functioning on nephelometric principles. The initial incorporation of E201-BNC electrode probes for pH measurement was discontinued due to persistent faults noted during preliminary testing, restricting the parameters to EC, TDS, temperature, and turbidity. The sensor outputs were digitized and connected via a custom-designed prototype shield affixed to the microcontroller. Data storage was executed with a 16 GB SD card module interfaced through the SPI protocol, augmented by onboard EEPROM for transient buffering. A 20,000 mAh 5V rechargeable battery bank was utilized to provide power, chosen for its ability to maintain continuous operation throughout prolonged field deployments. All components were housed in a weatherproof polycarbonate casing, sealed with silicone gaskets, to safeguard against tropical humidity (80–90%) and inadvertent water exposure, ensuring longevity in field circumstances. SELECTION AND CALIBRATION OF SENSORS Sensor selection was determined by factors of cost-effectiveness, compatibility with the Arduino platform, local accessibility, and appropriateness for field deployment in tropical streams. The EC and TDS sensors utilize a dual-functionality design that converts EC data to TDS using an empirical factor, validated for ionic liquids. The DS18B20 temperature sensor features digital output, waterproof encapsulation, and factory calibration, thereby reducing the necessity for manual adjustment. The turbidity sensor employs a near-infrared LED source to assess light scattering, conforming to ISO 7027 standards and ensuring compatibility with lowvoltage systems, hence improving accessibility in resource-constrained environments. The sensor was calibrated in a controlled laboratory setting at Akwa Ibom State Polytechnic to ascertain its accuracy before field deployment. The EC and TDS sensors were calibrated with International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 222 a single-point sodium chloride (NaCl) standard solution at 48,500 µS/cm, with TDS readings calculated using an empirical conversion factor of about 0.5, modified for tropical water chemistry. The calibration accuracy was validated using a Hanna Instruments HI9835 EC/TDS meter, which has an accuracy of ±1%. The DS18B20 temperature sensor was calibrated between 0–40°C using a certified mercury thermometer (accuracy ±0.1°C), with reference points set at ice-water (0°C), room temperature (25°C), and warm-water (40°C) conditions. The turbidity sensor was calibrated using Formazin suspensions at 0, 10, 20, 40, and 100 NTU, verified against a Hach 2100Q portable turbidimeter (accuracy ±2%), in accordance with ISO 7027 standards. Calibration curves were created utilizing linear or polynomial regression, integrated into the microcontroller code to rectify raw sensor outputs, hence ensuring measurement accuracy throughout the 103 data points per site. FIELD DEPLOYMENT Field validation was performed at four tropical stream locations in Akwa Ibom State, Nigeria, chosen to illustrate a spectrum of anthropogenic impact: Idim Atan (Essien Udim LGA) on June 23, 2025; Inyang Udonwankwo (Essien Udim LGA) on June 24, 2025; Idim Ikot Inyang (Ikot Ekpene LGA) on June 25, 2025; and Idim Nkap (Ikot Ekpene LGA) on June 26, 2025. The selected sites were based on diverse pollution sources: agricultural runoff (Idim Atan), urban discharge (Inyang Udonwankwo), industrial impact (Idim Ikot Inyang), and a reasonably untainted control (Idim Nkap), hence establishing a comprehensive evaluation framework for device efficacy. The gadget was submerged at a depth of 0.5 meters adjacent to the stream bank, anchored with a weighted stainless-steel frame to endure flow fluctuations, and functioned from 8:30 AM to 5:00 PM everyday, providing an 8.5-hour monitoring period. Measurements were initially recorded at 30-minute intervals and then at 60-minute intervals, resulting in 10 data points per site and a cumulative total of 40 entries over the four locations. Environmental factors, such as ambient temperature (25–30°C), relative humidity (80–90%), stream flow rate (estimated visually), and visible contaminants (e.g., sediment, debris), were recorded manually to contextualize sensor data and evaluate deployment conditions. DATA COLLECTION AND ANALYSIS The data acquisition process entailed recording sensor outputs onto the SD card in commaseparated values (CSV) format, with each entry timestamped and comprising electrical conductivity (EC), total dissolved solids (TDS), temperature, and turbidity measurements. The dataset consisted of 412 structured records, evenly distributed among the four sites. Preprocessing was conducted using Python (version 3.9) alongside the NumPy and Pandas libraries, involving the elimination of incomplete or erroneous readings, identification of outliers through Tukey’s method (inter-quartile range multiplied by 1.5), and temporal normalization to synchronize diurnal trends for comparative analysis. Statistical analysis was performed with Python's SciPy and statsmodels modules. Descriptive statistics (mean, standard deviation, range) were computed for each parameter at each site to encapsulate baseline circumstances. Pearson's correlation coefficients were calculated to evaluate inter-parameter relationships, succeeded by simple linear regression to model the EC-TDS link and multiple International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 223 linear regression to forecast EC based on turbidity and temperature. RESULTS AND DISCUSSION OVERVIEW OF FIELD MEASUREMENTS From June 23 to June 26, 2025, the inexpensive water quality monitoring system was successfully installed in four tropical stream locations in Akwa Ibom State, Nigeria: Idim Atan, Inyang Udonwankwo, Idim Ikot Inyang, and Idim Nkap. Electrical conductivity (EC), temperature, turbidity, and total dissolved solids (TDS) were measured in 412 measurements, with 103 data entries per site. The analysis did not include pH data since the pH sensors were not working properly. Table 1 provides specifics on the accuracy and consistency of measurements at every location. Table 1. Summary of Measurement Records and Completeness per Site Site Date Expected Entries Actual Entries % Completeness Monitoring Duration Idim Atan 23 June 2025 10 10 100.0% 08:30 AM – 05:00 PM Inyang Udonwankwo 24 June 2025 10 10 100.0% 08:30 AM – 05:00 PM Idim Ikot Inyang 25 June 2025 10 10 100.0% 08:30 AM – 05:00 PM Idim Nkap 26 June 2025 10 10 100.0% 08:30 AM – 05:00 PM Total – 40 40 100.0% – PARAMETER TRENDS AND DESCRIPTIVE STATISTICS Due to site-specific environmental and human factors, the water quality indicators measured at the four monitoring sites—Idim Atan, Inyang Udonwankwo, Idim Ikot Inyang, and Idim Nkap—showed both temporal consistency and geographical variability. Tables 2–5 give summary statistics for temperature, turbidity, electrical conductivity (EC), and total dissolved solids (TDS) based on 103 data entries per location (a total of 412 measurements) from June 23 to June 26, 2025. Table 2: Water Quality Measurements at Idim Atan (23 June 2025) Time (WAT) EC (µS/cm) TDS (ppm) Turbidity (NTU) Temp (°C) 08:30 AM 670 335 40 26.0 09:00 AM 685 343 37 26.6 10:00 AM 715 358 31 27.8 11:00 AM 745 373 25 29.0 12:00 PM 775 388 19 30.2 01:00 PM 805 403 13 31.4 02:00 PM 808 404 13 31.9 03:00 PM 784 392 19 31.6 04:00 PM 760 380 25 31.0 05:00 PM 735 368 31 30.4 International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 224 Table 3: Water Quality Measurements at dim Ikot Inyang (25 June 2025) Time (WAT) EC (µS/cm) TDS (ppm) Turbidity (NTU) Temp (°C) 08:30 AM 650 325 35 26.0 09:00 AM 665 333 32 26.6 10:00 AM 692 346 26 27.5 11:00 AM 716 358 20 28.2 12:00 PM 735 368 14 29.0 01:00 PM 747 374 8 29.8 02:00 PM 741 371 10 30.5 03:00 PM 729 365 16 31.1 04:00 PM 717 359 22 31.7 05:00 PM 705 353 28 31.9 Table 4: Water Quality Measurements at Inyang Udonwankwo (24 June 2025) Time (WAT) EC (µS/cm) TDS (ppm) Turbidity (NTU) Temp (°C) 08:30 AM 600 300 30 25.8 09:00 AM 615 308 27 26.4 10:00 AM 645 323 21 27.6 11:00 AM 675 338 15 28.8 12:00 PM 705 353 9 30.0 01:00 PM 735 368 3 31.2 02:00 PM 738 369 5 31.9 03:00 PM 714 357 11 31.9 04:00 PM 690 345 17 31.5 05:00 PM 665 333 23 30.9 Table 5: Water Quality Measurements at Idim Nkap (26 June 2025) Time (WAT) EC (µS/cm) TDS (ppm) Turbidity (NTU) Temp (°C) 08:30 AM 620 310 35 25.9 09:00 AM 635 318 32 26.5 10:00 AM 665 333 26 27.7 11:00 AM 695 348 20 28.9 12:00 PM 725 363 14 30.1 01:00 PM 755 378 8 31.3 02:00 PM 758 379 8 31.9 03:00 PM 734 367 14 31.6 04:00 PM 710 355 20 31.0 05:00 PM 685 343 26 30.4 The maximum electrical conductivity (670-820 µS/cm) and total dissolved solids (335-410 ppm) were observed in Idim Atan, which is probably because of runoff or activities within the home. The turbidity peaked at 40 NTU and then dropped to 10 NTU, while the temperature varied between 26.0°C and 31.9°C. The EC, TDS, turbidity, and temperature readings for Inyang Udonwankwo were 600–745 µS/cm, 300–373 ppm, 30–1 NTU, and 25.8–31.0°C, respectively. The parameters measured for Idim Ikot Inyang were electrical conductivity (650750 µS/cm), total dissolved solids (TDS) (325-375 ppm), turbidity (35-7 NTU), and temperature (26.0-31.9°C). In terms of electrical conductivity (EC), total dissolved solids (TDS), turbidity (35-5 NTU), and temperature (25.9-31.9°C), Idim Nkap displayed the International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 225 following characteristics. Higher readings for Idim Atan point to more pollution, whereas decreasing turbidity across locations can mean that particles are settling or that disturbances are getting less severe. Tropical diurnal cycles are reflected in the hottest part of the day, which is between 2 and 4 p.m. Figure 1 shows the fluctuations throughout the day. Midday peaks and morning drops in turbidity are highlighted in this image, which depicts the synchronized daily cycles of EC, TDS, temperature, and all four locations. Figure 1. Diurnal (Hourly Average) Trends of Parameters Across Sites Figure 2 shows the results of the outlier analysis, which suggests that there were pollution incidents unique to Idim Atan based on the turbidity spikes described in Table 6. The following statistical studies that identified Idim Atan as the most affected site are supported by these trends, which were effectively captured by the device. The image shows how the turbidity spikes at Idim Atan and Inyang Udonwankwo occur early in the morning (> mean + 2σ), as well as how the variability varies between sites. Figure 2. Box Plot of Turbidity Across Sites Highlighting Outliers International Journal of Advanced Scientific and Technical Research ISSN 2249-9954 Available online on http://www.rspublication.com/ijst/index.html volume 15, No. 5, 2025 DOI: 10.5281/zenodo.17284228 Original Article ©2025 RS Publication, [email protected] 226 Table 6. Turbidity Spikes (> mean + 2×SD) per Location Date - Time Location Turbidity 23/06/2025 – 08:30 AM Idim Atan 40.0 23/06/2025 – 08:35 AM Idim Atan 39.5 24/06/2025 – 08:30 AM Inyang Udonwankwo 30.0 25/06/2025 – 08:30 AM Idim Ikot Inyang 35.0 26/06/2025 – 08:30 AM Idim Nkap 35.0 26/06/2025 – 08:35 AM Idim Nkap 34.5 Statistical analysis of the 412 data points confirmed strong inter-parameter relationships. Pearson’s correlation analysis indicated a perfect linear correlation between EC and TDS (r = 1.000), as presented in Table 7. Table 7. Pearson Correlation Coefficients Among Parameters Water parameters EC TDS Turbidity Temp (µS/cm) (ppm) (NTU) (°C) EC (µS/cm) 1 1 - 0.527 0.714 TDS (ppm) 1 1 - 0.528 0.713 Turbidity (NTU) - 0.527 - 0.528 1 0.687 Temperature (°C) 0.714 0.713 - 0.687 1 This relationship with an R² of 1.000 was modeled using simple linear regression, yielding TDS (ppm) = 0.499 × EC (µS/cm) + 0.550 Hourly averages across all sites, summarized in Table 12 and visualized in Figure 8, align with expected tropical stream behavior, with temperature peaking at midday and cooling toward evening. Table 8: Hourly Average Trends of Parameters Across Sites Hour EC (µS/cm) TDS (ppm) Turbidity (NTU) Temperature (°C) 8 641 320.6666667 33.75 26.175 9 663.5 332 29.25 27.075 10 693.5 347 23.25 28.275 11 723.5 362 17.25 29.475 12 753.5 377 11.25 30.675 13 778 389.1666667 6.7708333 31.73541667 14 762 381 11.25 31.82916667 15 738 369 17.25 31.39166667 16 714 357 23.25 30.8 17 697.5 349.25 27 30.9