UVR prediction over Lagos Ibadan and New Richmond for Public Health and Renewable Energy Applications

doi:10.21203/rs.3.rs-4713224/v1

Download PDF

Research Article

UVR prediction over Lagos Ibadan and New Richmond for Public Health and Renewable Energy Applications

https://doi.org/10.21203/rs.3.rs-4713224/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Ultraviolet radiation (UVR) is a significant environmental factor influencing various biological and chemical processes, including photosynthesis in plants, vitamin D synthesis in humans, and microbial sterilization. However, excessive UVR exposure can lead to adverse effects such as skin cancer and DNA damage. This study applies the ARIMA (AutoRegressive Integrated Moving Average) model to predict UVR levels over Lagos and Ibadan in Nigeria, and New Richmond in the United States, utilizing a 21-year dataset spanning from 2000 to 2020. By analysing the autocorrelation function (ACF) and partial autocorrelation function (PACF) with a significance level of 0.25, the stationarity and appropriate parameters for the ARIMA model were identified. The model is then used to predict daily UVR values for the last day of each month from January 2021 to December 2022. Results indicate that the ARIMA model effectively captures the temporal patterns in UVR data, with validation metrics such as RMSE, MAE, and MAPE confirming its predictive accuracy. This predictive capability can inform public health advisories, agricultural practices, and environmental management, emphasizing the need for ongoing monitoring and prediction of UVR levels to mitigate potential health risks.

Ultraviolet Radiation (UVR)

ARIMA Model

Time Series Prediction

Lagos

Ibadan

New Richmond

ACF

PACF

RMSE

MAE

MAPE

UV Exposure

Environmental Health

Climate Data Analysis

Predictive Modelling.

Ultraviolet radiation (UVR) is a type of electromagnetic radiation with wavelengths ranging from 100 to 400 nanometres, falling between X-rays and visible light on the electromagnetic spectrum. It is a natural component of sunlight and can be divided into three types: UVA (320–400 nm), UVB (290–320 nm), and UVC (100–290 nm) [1, 2]. While UVC is mostly absorbed by the Earth's atmosphere, UVA and UVB reach the surface, influencing a range of biological and environmental processes. UVR has both beneficial and detrimental effects. On the positive side, it plays a crucial role in the photosynthetic processes of plants, the production of vitamin D in humans, and the sterilization of water and surfaces due to its germicidal properties [3, 4, 5]. However, excessive exposure to UVR can lead to harmful consequences such as skin burns, increased risk of skin cancer [5, 6], and DNA mutations [6], which can adversely affect both plant and animal life. Understanding and predicting UVR levels are essential for several applications, particularly in public health and renewable energy sectors. Accurate UVR predictions can help in planning and implementing protective measures against harmful UV exposure, thereby reducing the incidence of skin-related diseases and improving overall public health. In the renewable energy sector, UVR data is valuable for optimizing the performance of solar energy systems, as UV radiation is a component of solar irradiance that influences the efficiency of photovoltaic cells.

This study focuses on the application of the ARIMA (AutoRegressive Integrated Moving Average) model for predicting UVR levels over three distinct locations: Lagos and Ibadan in Nigeria, and New Richmond in the United States. Utilizing a comprehensive dataset spanning 21 years (2000–2020) from meteoblue.com, this research aims to develop accurate predictive models by analysing the autocorrelation function (ACF) and partial autocorrelation function (PACF) of the UVR data over Lagos and Ibadan in Nigeria, and New Richmond in the United States, thereby contributing to public health and renewable energy applications. However, the objectives of this paper are to: collect and preprocess a 21-year dataset (2000–2020) of UV radiation measurements from Lagos, Ibadan, and New Richmond, analyse the trends, seasonality, and stationarity of UVR data using statistical methods, develop an ARIMA model for each study location based on the identified parameters from ACF and PACF analysis. As well as exploring the potential applications of UVR predictions in optimizing the performance and efficiency of solar energy systems. The ARIMA model, known for its effectiveness in time series forecasting [7, 8, 9], is employed to predict daily UVR values for the last day of each month from January 2021 to December 2022. By providing reliable UVR forecasts, this research seeks to contribute to the enhancement of public health strategies and the advancement of renewable energy technologies. The findings can inform policymakers, healthcare providers, and energy planners, enabling them to make data-driven decisions that promote safety, efficiency, and sustainability.

Excessive exposure to UVR can have detrimental effects, such as skin cancer, DNA damage, and other health issues. These literatures review explores previous studies on UVR, its impacts, and the application of ARIMA models for time series prediction in various domains, including environmental monitoring and public health. Several studies have highlighted the public health implications of UVR exposure. [3] discussed human skin pigmentation as an adaptation to UVR, emphasizing its protective role against DNA damage and skin cancer. Similarly, [1] examined the molecular responses of human skin to repetitive UV exposure, illustrating how both UVA and UVB contribute to skin aging and carcinogenesis. The detrimental effects of UVR on health underscore the need for accurate prediction models to inform public health advisories and protective measures. ARIMA models have been effectively applied in various fields, including environmental monitoring, economics, and engineering. [7] utilized ARIMA models to forecast particulate matter concentrations in Taiyuan, China, demonstrating the model's capability to handle environmental data with seasonal fluctuations. Similarly, [9] applied ARIMA models to predict water levels in the Yangtze River, aiding in flood prevention and water resource management. These studies highlight the versatility and effectiveness of ARIMA models in handling complex temporal data.

This study employs a robust methodology to predict ultraviolet radiation (UVR) levels over Lagos, Ibadan, and New Richmond, with applications in public health and renewable energy. The methodology is structured into several key phases: data collection, data preprocessing, model development, and model validation.

Study Locations

Below are the geographical details of each location such as country, time zones, latitudes, longitudes, above the sea level measurements and the date range of these locations are listed Table 1 below:

Table 1

The Geographical Details of The Selected Locations
Locations	Country	Time Zone	latitude (m)	longitude (m)	Height (m)	Duration
Ibadan [13]	Nigeria	+ 1:00 GMT	7.3767	3.9398	215.2770	2000–2020
Lagos [14]	Nigeria	+ 1:00 GMT	6.4654	3.4065	10.8770	2000–2020
New Richmond	U. S.	-5:00 GMT	38.9227	-84.3750	227.2680	2000–2020

Data Collection

UVR data for Lagos, Ibadan, and New Richmond is collected from meteoblue.com, using observations. The data spans for twenty-one years to capture seasonal and annual variations in UVR levels, which is the mean daily values of hourly data of the last days of every month from the 31st of January 2000, to the 31st of December, 2020. However, Fig. 1 below presents graphical design of prediction process.

Additionally, meteorological data such as temperature, humidity, and cloud cover, which influence UVR, is also collected from the same locations to enhance the accuracy of predictive model [8].

Data Preprocessing

Data preprocessing involves several steps to ensure the accuracy and reliability of the input data:

Data Cleaning: This step involves handling missing values and outliers. Missing data points are imputed using statistical techniques such as mean imputation and interpolation [7].

Normalization: To facilitate effective modelling, the data is normalized to a standard scale. This helps in mitigating the effects of different magnitudes of variables and ensures better convergence of the model [9].

Stationarity Testing: Since ARIMA models require stationary data, the dataset is tested for stationarity using the Augmented Dickey-Fuller (ADF) test. If the data is found to be non-stationary, differencing is applied to achieve stationarity [10].

Model Development

In the context of UVR prediction, ARIMA models can be particularly useful due to their ability to capture temporal patterns and seasonal variations. The application of ARIMA models to UVR data involves several steps, including data preprocessing, stationarity analysis using ACF and PACF, model fitting, and validation [10]. By accurately predicting UVR levels, these models can inform public health advisories and optimize the performance of solar energy systems. The ARIMA model is selected for its robustness in handling time series data with temporal patterns and seasonality. The development of the ARIMA model involves the following steps:

Identification: The autocorrelation function (ACF) and partial autocorrelation function (PACF) are used to determine the appropriate order of the ARIMA model [11], denoted as ARIMA (p, d, q), where:

(p) is the number of lag observations included in the model (autoregressive part).

(d) is the number of times that the raw observations are differenced to achieve stationarity.

(q) is the size of the moving average window.

ARIMA model can be expressed as a combination of autoregressive (AR), moving average (MA) parameters and lagging error:

AR = $\:{V}_{t}=\:{\phi\:}_{1{V}_{t-1}\:}+\:{\phi\:}_{2}{V}_{t-2}---\:+\:{\phi\:}_{p}{V}_{t-p}+\:{\mu\:}_{t}\dots\:\dots\:\dots\:\left(1\right)$

MA = $\:{\mu\:}_{t}=\:{\epsilon\:}_{t}-{\theta\:}_{1}{\epsilon\:}_{t-1}-\:{\theta\:}_{2}{V}_{t-2}---\:+\:{\theta\:}_{q}{\epsilon\:}_{t-q}\dots\:\dots\:\dots\:\dots\:(2$)

Therefore, the ARIMA model is written as:

$$\:{V}_{t}=\:\mu\:+\:{\phi\:}_{1}{V}_{t-1}---\:+\:{\phi\:}_{p}{V}_{t-p}-\:{\theta\:}_{1}{\epsilon\:}_{t-1}---\:-{\theta\:}_{q}{\epsilon\:}_{t-q}\dots\:\dots\:\dots\:\left(3\right)$$

Where $\:\mu\:\:$is the constant.

$\:{V}_{t}$ is the simple value.

$\:\epsilon\:$ is the sequence of white noise.

$\:\theta\:$ (i = 1, 2, 3, …q) is parameter of moving average.

$\:\phi\:$ (i = 1, 2, 3, …p) is the parameter of autoregression.

$\:{\theta\:}_{q}{\epsilon\:}_{t-q}$ is the lagged error.

$\:{\phi\:}_{p}{V}_{t-p}$ are the lagged values of V. [8, 10]

Estimation: Using the identified order, the parameters of the ARIMA model are estimated through maximum likelihood estimation [10].

Diagnostic Checking: Residual analysis is performed to ensure that the residuals (errors) of the model are white noise, indicating a good fit. This involves checking the residuals for autocorrelation and ensuring they are normally distributed.

Model Validation

To assess the predictive model performance, the data is divided into training and testing sets. The training set is used to build the model, while the testing set evaluates its accuracy. Key performance metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE) are calculated to quantify the model's accuracy [7].

Implementation and Applications

After validation, the ARIMA model is used to forecast UVR levels for the selected locations. These forecasts are then applied to public health and renewable energy applications:

Public Health Applications: Predicted UVR levels are used to issue advisories and warnings to the public to mitigate health risks associated with UV exposure. This includes recommendations for sunscreen use, protective clothing, and avoiding outdoor activities during peak UV hours [1].

Renewable Energy Applications: In the context of solar energy, accurate UVR predictions help optimize the performance of photovoltaic systems. By forecasting UVR, solar energy production can be better managed, improving the efficiency and reliability of solar power generation [12].

Hence, the following steps were taken to apply the ARIMA model (Box – Jenkins model) to the dataset.

Step 1: standardization of the dataset
Step 2: trend, stationarity and seasonal analysis of the dataset.
Step 3: reconstruction of the dataset to have mean daily data from the 31st of January 2000 to the 31st of December 2020.
Step 4: fit an ARIMA model to the dataset as well as the reconstructed dataset.

Step 5: the prediction of the mean daily values of the last days of each month between January 31st 2021 and December 31st 2022 using the Python programming language.

Python Programming Language

Spectrum analysis and prediction were programmed on the Python environment. The spectrum analysis was carried out by wavelet analysis, Continuous Wavelet Transform (CWT) specifically and the prediction was carried out with the help of the ARIMA Model also known as the Box-Jeckins Model. Therefore, the dataset was analysed in the Python environment using SPYDER specifically through the help of Python libraries such as pycwt – using Morlet wavelet function (Morl6) for wavelet functions, pandas – to call the dataset file from its location on the system or cloud by coping the path of the file, NumPy – for numerical functions and others, matplotlib.pyplot – for visualisation. Equally, ARIMA and pmdarima were improved libraries for predictive functions. The dataset was cleaned, and no missing datum was found, after which the data description, data info, and data head were also derived.

Figures 2 and 3 depict the plots of PACF and ACF of UVR over Lagos, Figs. 4 and 5 represent PACF and ACF of UVR over Ibadan while the PACF and ACF of UVR over New Richmond are tagged as Figs. 6 and 7.

In Figs. 2, 4 and 6, only the PACFs are tending towards zero with an increase in lags which means the UV radiation datasets over Lagos, Ibadan and New Richmond are stationary, while the ACFs of Figs. 3, 5 and 7 are not tending to zero while the lags increase. That is, the ACF cannot remove the non-stationarity in the UV radiation over Lagos, Ibadan and New Richmond.

Daily mean predicted values of UV radiation over the study locations on every last day of each month from January 2021 to December 2022, which is from 31st of January 2021 to 31st of December 2022 of UVR over the study locations (Lagos, Ibadan and New Richmond) are featured in Figs. 8, 9 and 10. However, UVR ARIMA fitted for Lagos, Ibadan and New Richmond are represented by Figs. 11, 12 and 13 respectively.

The Results of Partial Correlation PACF and Autocorrelation ACF

The results images are available in the Figures carousel.

PACF and ACF Discussion

The predictive models, which will be applied in the arrangement are selected about the stationarity of the Autocorrelation function (ACF) and Partial Correlation function (PACF) when the stationarity is achieved, followed by future prediction [11, 12]. The values of p and q can be determined by the time series stationarity of ACF and PACF results. A dataset is stationary when ACF is increased relative to the lag and rapidly tending to zero (0), and non-stationary when the ACF is increased in relation to the lag and does not tend to zero (0) [8].

Mean Daily Prediction Discussion

Figures 8, 9 and 10 show the graphs of the mean daily values of UV radiation over Lagos, Ibadan and New Richmond respectively from 2000 to 2022 which comprise observed mean daily values and predicted mean daily (the mean value of each last day of every month in the study period – 2000–2020) values of the UV radiation over the study areas. In Figs. 8, 9 and 10, the blues colour lines represent the observed mean daily values of UV radiation over the study areas while the orange colour lines represent the predicted mean daily values of the UV radiation over the study areas.

ARIMA Fitted

The graph images are available in the Figures carousel.

The discussion of ARIMA Fitted results

The graphs of ARIMA fitted values of UVR in Lagos, Ibadan and New Richmond are shown in Figs. 10, 12 and 13 respectively. ARIMA fitted was employed to evaluate errors in the predicted values through the help of Root Mean Square Error (RMSE), Mean Average Error (MAE), and Mean Average Percentage Error (MAPE) where the original (raw) data were presented in green colour and the red colour represented the predicted values over each study locations. The ARIMA-fitted values of UV radiation in the study locations – Lagos, Ibadan and New Richmond, showed that the original (raw) data and the predicted data are correlated.

Discussion of Results

The results obtained from the ARIMA modelling of ultraviolet radiation (UVR) levels over Lagos, Ibadan, and New Richmond provide valuable insights into the temporal variations and predictive capabilities of the models. This discussion focuses on the accuracy of the forecasts, the implications for public health and renewable energy sectors, and potential avenues for further research.

Accuracy of Forecasts: The ARIMA models developed for predicting UVR levels demonstrated robust performance across the three locations. Key metrics such as Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE) were used to evaluate the accuracy of the forecasts. Generally, the models achieved low error rates, indicating a good fit to the historical UVR data. For instance, the MAPE values ranged within acceptable limits, suggesting that the models are reliable for forecasting UVR levels in these regions.
Implications for Public Health: Accurate UVR predictions are crucial for public health initiatives aimed at minimizing the adverse effects of UV exposure on human health. With reliable forecasts, authorities can issue timely advisories to the public about UV intensity and recommend appropriate measures such as using sunscreen, wearing protective clothing, and avoiding outdoor activities during peak UV hours. The findings of study underscore the importance of integrating UVR forecasting into public health policies to mitigate risks associated with skin cancers, sunburns, and other UV-related health conditions.
Implications for Renewable Energy: In the context of renewable energy, precise UVR predictions are essential for optimizing the performance of solar energy systems. By forecasting UVR levels, solar power generation can be better managed, leading to increased efficiency and reliability of solar panels. This application is particularly relevant in regions like Lagos, Ibadan, and New Richmond, where solar energy potential is significant but highly dependent on local weather conditions, including UVR intensity.

Limitations and Future Directions

Despite the promising results, several limitations warrant consideration for future research. Firstly, the ARIMA models assume stationarity and may not capture abrupt changes or nonlinear trends in UVR levels. Incorporating more advanced time series models or machine learning techniques could enhance predictive accuracy, especially in capturing complex interactions between meteorological variables and UVR. Furthermore, expanding the geographical scope beyond these three locations could provide a broader understanding of UVR patterns across different climatic zones and latitudes. Additionally, integrating real-time data from advanced satellite platforms or ground-based sensors could improve the temporal resolution and reliability of UVR forecasts.

This study applied ARIMA modelling to forecast ultraviolet radiation (UVR) levels over Lagos, Ibadan, and New Richmond, highlighting its significance for public health and renewable energy applications. The results indicate that ARIMA models effectively predict UVR levels with reasonable accuracy, as evidenced by low error metrics such as Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE). Furthermore, leveraging ARIMA modelling for UVR prediction offers substantial benefits for both public health and renewable energy sectors. Continued advancements in modelling techniques and data integration are crucial for enhancing the reliability and applicability of UVR forecasts in mitigating health risks and promoting sustainable energy solutions. These forecasts are vital for proactive public health measures and optimizing solar energy utilization in these regions. However, this discussion highlights the importance of interdisciplinary approaches in leveraging meteorological data for practical applications, ultimately contributing to sustainable development and improved quality of life in urban and semi-urban settings.

Recommendations

Based on the findings of this research, the following recommendations are made:

Enhanced Monitoring and Forecasting: Implement enhanced monitoring systems for UVR levels using advanced satellite data and ground-based sensors. This will improve the temporal resolution and accuracy of UVR forecasts, aiding in better public health advisories and solar energy management.

Integration with Public Health Policies: Integrate UVR forecasting into public health policies and strategies to mitigate health risks associated with UV exposure. Timely dissemination of UVR forecasts can empower individuals to take preventive measures such as sunscreen use and limiting outdoor activities during peak UV hours.

Research Expansion: Expand research efforts to include more geographical locations and diverse climatic conditions. This broader scope will provide a comprehensive understanding of UVR variability across different regions, enhancing the generalizability of predictive models.

Model Refinement: Explore advanced modelling techniques beyond ARIMA, such as machine learning algorithms, to capture nonlinearities and complex interactions influencing UVR levels. This approach can further improve the accuracy and robustness of UVR forecasts.

Education and Awareness: Promote public education and awareness campaigns on the health effects of UV radiation and the importance of UVR forecasting. Empowering communities with knowledge can encourage proactive health behaviours and reduce the incidence of UV-related illnesses.

Policy Support for Solar Energy: Support policies that incentivize the adoption of solar energy technologies based on UVR forecasts. By optimizing solar energy production, communities can reduce reliance on non-renewable energy sources, contributing to environmental sustainability.

Author Contribution

Author ContributionsSherifdeen M. Oyedokun : Conceptualization, Methodology, Data Curation, Writing - Original Draft Preparation, Visualization.Gbadebo I. Olatona: Supervision, Project Administration, Writing - Review & Editing, Formal Analysis.Shuayb A. Adisa: Data Collection, Software, Validation, Investigation.All authors have read and agreed to the published version of the manuscript.

Data Availability

Sequence data which support the results of this research was retrieved from meteoblue.com and saved on the link:

Choi, W., Miyamura, Y., Wolber, R., Smuda, C., Reinhold, W., Liu, H., Kolbe, L., & Hearing, V. J. (2010). Regulation of human skin pigmentation in situ by repetitive UV exposure: Molecular characterization of responses to UVA and/or UVB. Journal of Investigative Dermatology, 130(7), 1685–1696. DOI: 10.1038/jid.2010.5.
Solano, F. (2020). Photoprotection and Skin Pigmentation: Melanin-Related Molecules and Some Other New Agents Obtained from Natural Sources. Molecules, 25(7), 1537. DOI: 10.3390/molecules25071537.
Jablonski, N. G., & Chaplin, G. (2010). Human skin pigmentation as an adaptation to UV radiation. Proc Natl Acad Sci U S A, 107(20), 8962–8968. DOI: 10.1073/pnas.0914628107.
Jablonski, N. G. (2004). The Evolution of Human Skin and Skin Color. Annual Review of Anthropology, 33, 585–623. http://www.jstor.org/stable/25064866.
Miyamura, Y., Coelho, S. G., Wolber, R., Miller, S. A., Wakamatsu, K., Zmudzka, B. Z., Ito, S., Smuda, C., Passeron, T., Choi, W., Batzer, J., Yamaguchi, Y., Beer, J. Z., & Hearing, V. J. (2006). Regulation of human skin pigmentation and responses to ultraviolet radiation. Pigment Cell Research, 20(1), 3–13. DOI: 10.1111/j.1600-0749.2006.00358.x.
Tadokoro T, Yamaguchi Y, Batzer J, Coelho SG, Zmudzka BZ, Miller SA, Wolber R, Beer JZ, Hearing VJ. Mechanisms of skin tanning in different racial/ethnic groups in response to ultraviolet radiation. J Invest Dermatol. 2005;124(6):1326-32. DOI: 10.1111/j.0022-202X.2005.23760.x. PMID: 15955111.
Zhang, H., Zhang, S., Wang, P., Qin, Y., & Wang, H. (2017). Forecasting of PM10 time series using wavelet analysis and wavelet-ARMA model in Taiyuan, China. Journal of the Air & Waste Management Association, ISSN: 2162–2906. DOI: 10.1080/10962247.2017.1292968.
Lihua, N., Xiaorong, C., & Qian, H. (2010). ARIMA model for traffic flow prediction based on wavelet analysis. In The 2nd International Conference on Information Science and Engineering (pp. 1028–1031). IEEE.
Yu, Z. G., Lei, Z., & Jiang, F. (2017). ARIMA Modelling and Forecasting of Water Level in the Middle Reach of the Yangtze River. In 2017 4th International Conference on Transportation Information and Safety (ICTIS), Banff, Canada. IEEE. DOI: 10.1109/ICTIS.2017.8047799.
Box, G. E. P., Jenkins, G. M., Reinsel, G. C., & Ljung, G. M. (2015). Time Series Analysis: Forecasting and Control. John Wiley & Sons.
Zhen L., and Y. A. Hassan, (2006). Wavelet autocorrelation identification of the turbulent flowmulti-scales for drag reduction process in micro-bubbly flows. Chemical Engineering Science 61. Elsevier. Page: 7107–7114. DOI: DOI:10.1016/j.ces.2006.07.031.
Tian, Y. (2017). Stock forecasting method based on wavelet analysis and ARIMA-SVR model. In 2017 3rd International Conference on Information Management (pp. 102–106). DOI: 10.1109/INFOMAN.2017.8120305.
LatLong.net. (2021). Ibadan, Nigeria. Retrieved from https://www.latlong.net/place/ibadan-nigeria-1748.html.
LatLong.net. (2021). Lagos, Nigeria. Retrieved from https://www.latlong.net/place/lagos-nigeria-2286.html

No competing interests reported.

Download PDF

Editorial decision: Revision requested
24 Jul, 2024
Editor assigned by journal
23 Jul, 2024
Submission checks completed at journal
23 Jul, 2024
First submitted to journal
09 Jul, 2024

You are reading this latest preprint version

UVR prediction over Lagos Ibadan and New Richmond for Public Health and Renewable Energy Applications

Status:

Version 1

Abstract

Figures

Introduction

METHODOLOGY

Study Locations

Data Collection

Data Preprocessing

Model Development

AR = \(\:{V}_{t}=\:{\phi\:}_{1{V}_{t-1}\:}+\:{\phi\:}_{2}{V}_{t-2}---\:+\:{\phi\:}_{p}{V}_{t-p}+\:{\mu\:}_{t}\dots\:\dots\:\dots\:\left(1\right)\)

MA = \(\:{\mu\:}_{t}=\:{\epsilon\:}_{t}-{\theta\:}_{1}{\epsilon\:}_{t-1}-\:{\theta\:}_{2}{V}_{t-2}---\:+\:{\theta\:}_{q}{\epsilon\:}_{t-q}\dots\:\dots\:\dots\:\dots\:(2\))

Model Validation

Implementation and Applications

Python Programming Language

RESULT AND DISCUSSION

The Results of Partial Correlation PACF and Autocorrelation ACF

PACF and ACF Discussion

Mean Daily Prediction Discussion

ARIMA Fitted

The discussion of ARIMA Fitted results

Discussion of Results

Limitations and Future Directions

Conclusion

Recommendations

Declarations

Author Contribution

Data Availability

References

Additional Declarations

Status:

Version 1