Forecasting lake evaporation under a changing climate with an integrated artificial neural network model: A case study Lake Nasser, Egypt

https://doi.org/10.1016/j.jafrearsci.2021.104191Get rights and content

Highlights

  • This paper developed Artificial Neural Network model to assess Lake Nasser evaporation.

  • The developed model is calibrated and validated for Lake Nasser.

  • The paper assessed the impact of climate change using the developed model on Lake Nasser.

  • The model could be tested for other dry regions with a good degree of trust.

  • The most influential meteorological parameters on Lake Nasser projected evaporation.

Abstract

Understanding lake evaporation and the climate change role in evaporation is paramount for any water resources management system. The prediction of the climate's future changes is a very important step in planning lake future management decisions. This study analyzed Lake Nasser's evaporation in southern Egypt. Meteorological parameters were compiled from Aswan meteorological station near Lake Nasser. Also, CORDEX predicted climatological parameters from 2021 to 2050 were collected. Lake Nasser's evaporation prediction model using artificial neural networks technique was built. Statistics were calculated in the calibration and validation stages to find the most adequate model of the Lake evaporation calculation. The predictions of future evaporation were extracted from the model using predicted climatological data from CORDEX regional climate models. Trend analysis was done to assure the impacts of climate change on the lake. ANN model was developed and implemented successfully on Lake Nasser, which could be used to handle evaporation calculation over Lake Nasser. ANN model with training algorithm with 2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22 neurons with 5 input variables were tested to find the best model evaporation estimation of the lake with the least number of neurons. ANN model with training algorithm with 20 neurons with 5 input variables performed the best for evaporation estimation of the lake. According to predicted climate data, about a 2% increase in lake Nasser evaporation could be predicted in the year 2050. It was noticed that the maximum predicted values of evaporation are in July to August months with a range from 7.04 mm/day to 9.64 mm/day. The peak value, the outlier of maximum evaporation, is about 11.16 mm/day. The minimum predicted values are in December to January months with a range from 3.50 mm/day to 6.81 mm/day. Trend analysis showed that the predicted climatological parameters are slightly higher than historical records.

Introduction

Our planet is changing due to climate variations. These variations are affected mainly by anthropogenic interventions. Climate variability refers to statistical deviations of climate over a specific period of time. But climate change denotes a considerable climatological shift. Water is the main stem of living on the Earth. It is essential to most anthropogenic activities. Water plays a complex and multifunctional role in human life and society (El-Mahdy, 2014). The Nile River is the prime source of fresh water in Egypt. Most of Egypt's area is defined as semi-arid to arid area, with too little rainfall and underground water (Entz, 1976). After building High Aswan Dam (HAD) as a multipurpose project, Lake Nasser the long-term storage reservoir was created. The lake is born in both Egypt and Sudan between latitudes 20° 27' N and 23° 58' N, and longitudes 30° 07' E and 33°15 E (Badawy, 2009). Its maximum surface area is about 6500 Km2 and its maximum capacity is 162 BCM at elevation 182m above mean sea level (amsl); its mean depth 25m. The Lake is about 500 Km long (about 350 Km in Egypt and about 150 Km in Sudan), while the maximum width is reaching about 35 Km (El-Mahdy et al., 2018). The lake is considered the main stem of the Egyptian economy. Fig. 1 shows a site map of Lake Nasser. Its importance to Egypt cannot be denied. The arable lands around the Lake region represent the future food source for Egypt.

If the Nile River was not in Egypt, the whole country may not be there. The population of Egypt is predicted to be 124 million within 30 years. Although HAD was finished in 1970, but the surrounding agricultural land around the area is not exceeding 4% of the available area. The main restriction in front of agriculture extension is the available water, not the arable soil. Climate change is a very challenging and complex problem facing all the researchers in the forecasting of future hydrologic cycles worldwide. Current predictions of climate change's impact on hydrologic cycles are largely pessimistic in many parts of the world. In general, evaporation is the phenomenon by which a substance is converted from the liquid phase into the vapor phase (Monteith, 1965). Evaporation is pertinent to different disciplines of various scales. It reduces the available water quantity and quality because the evaporation process takes pure water only (Monteith, 1981). Climate change proves to impact the continent of Africa as shown in Fig. 2. For all the above-mentioned reasons, better quantification of Lake Nasser evaporation, in addition to the impact of climate change on it, is a must. In this study, the environmental influences, incorporating meteorological and hydrological records over the Lake, were taken into consideration. These records are used to explore the evaporation and to predict the climate change impact on the lake hydrology. Most of the evaporation prediction models used worldwide depend on the basics of energy balance, aerodynamic, a combination of both, or empirical formulas. The equations, that were often used for calculating evaporation, should undergo careful calibrations, so they have constraints on a global scale. Taking these constraints into account, the traditionally used evaporation techniques need improvements using lesser complex techniques. The impact of climate change on lakes is studied in many papers (Beyene, 2016; Erler et al., 2019; Fontana et al., 2019). The hydrologists used many methods to calculate hydrological phenomena, one of them is the Artificial Neural Network (Asadi et al., 2019; Humphrey et al., 2016; Pradhan et al., 2020; Zema et al., 2020). Many papers studied the impact of meteorological parameters on Lake Nasser (Khedr et al., 2013, 2014). The objective of the study is to develop an accurate evaporation model for Lake Nasser and to apply it to predict the future evaporation of the lake for better planning and management.

Section snippets

Evaporation prediction methods

Evaporation prediction Methods are often used data that is not always available for all regions. The evaporation rate from the free water surface depends on meteorological parameters of the overlying air. It also depends on the balance of energy between the air and water, and the quantity of stock energy inside the body of water. Due to the obscurity of evaporation gauging, the calculation models are the only available alternatives. The number of suggested models to quantify evaporation from

Artificial neural network (ANN)

ANN is a largely parallel-distributed data processing system. It imitates the process of the neuron system in the brain. Neurons' network in the human brain is responsible for their ability to learn. This considerable property is the base of machine learning in ANN (McCulloch and Pitts, 1943). (ANN) is also a computational method devised in the researches of the nervous system of living organisms. ANNs introduce perfect mathematical models simulating complex systems. The most significant

Hydrological data

The evaporation data used in this study for each month from January 1997 to July 2007 taken from the Ministry of Water Resources and Irrigation reports.

Meteorological data

The meteorological data used in this study for each month from January 1997 to July 2007 measured at Aswan meteorological station are as follows:

  • 1.

    Mean monthly air temperature.

  • 2.

    Mean monthly wind speed at a height of 2 m.

  • 3.

    Mean monthly sea level pressure.

  • 4.

    Mean monthly relative humidity.

  • 5.

    Mean monthly solar radiation.

Predicted meteorological data

Global Climate Models (GCM) can

Methods

MATLABFX is a highly qualified software for mathematical modeling. It is founded and originally evolved by mathematical programmers (Moler, 2004). MATLAB is devoted to calculations of matrices. The matrix is a mathematical term that denotes an array of numbers as follows:A=(1234)

Results and discussion

In this study, the ideal number of hidden layer neurons in the ANN model was specified by trial and error procedure. The procedure launched initially with two neurons and increased to 22 with a step size of two. For each trial, the ANN model was trained in a bulk mode to idealize the MSE at the output neurons. To examine the over-fitting, cross-validation was applied through the training process. Always, the available modeling data is divided into three datasets (training, validation, and

Conclusions

Lake Nasser's evaporation prediction model using artificial neural networks technique was built. Statistical analysis has been done in the calibration and validation stages to find the ideal model of the Lake evaporation calculation. The predictions of future evaporation were extracted from the model using predicted climatological data. The data from CORDEX regional climate models with two emission scenarios; scenario RCP 4.5 and scenario RCP 8.5 were used. Trend analysis was done to assure the

Data availability statement

Some or all data, models, or code generated or used during the study are available from the corresponding author by request.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References (71)

  • S. Palani et al.

    An Ann application for water quality forecasting

    Mar. Pollut. Bull.

    (2008)
  • P. Pradhan et al.

    Evaluation of soil and water assessment tool and artificial neural network models for hydrologic simulation in different climatic regions of asia

    Sci. Total Environ.

    (2020)
  • J.C. Refsgaard

    Parameterisation, calibration and validation of distributed hydrological models

    J. Hydrol.

    (1997)
  • B. Shi et al.

    Applying high-frequency surrogate measurements and a wavelet-Ann model to provide early warnings of rapid surface water quality anomalies

    Sci. Total Environ.

    (2018)
  • D.A. Zema et al.

    Predicting the hydrological response of a forest after wildfire and soil treatments using an Artificial Neural Network

    Comput. Electron. Agric.

    (2020)
  • J.-R. Zhang et al.

    A hybrid particle swarm optimization–back-propagation algorithm for feedforward neural network training

    Appl. Math. Comput.

    (2007)
  • H. Al-Barqawi et al.

    Infrastructure management: integrated Ahp/Ann model to evaluate municipal water mains' performance

    J. Infrastruct. Syst.

    (2008)
  • H. Asadi et al.

    Rainfall-runoff modelling using hydrological connectivity index and artificial neural network approach

    Water

    (2019)
  • H. Badawy

    Effect of expected climate changes on evaporation losses from aswan high Dam reservoir (ahdr)

  • E.A. Beyene

    Impact of climate change and anthropogenic activities on livelihood of fishing community around Lake Tana, Ethiopia

    Ec Agriculture

    (2016)
  • T. Beyene et al.

    Hydrologic impacts of climate change on the Nile River Basin: implications of the 2007 Ipcc scenarios

    Climatic Change

    (2010)
  • A.M. Binley et al.

    Changing responses in hydrology: assessing the uncertainty in physically based model predictions

    Water Resour. Res.

    (1991)
  • A.K. Biswas

    Lake Nasser: alleviating the impacts of climate fluctuations and change

  • P. Block et al.

    Economic analysis of large-scale upstream river basin development on the Blue Nile in Ethiopia considering transient conditions, climate variability, and climate change

    J. Water Resour. Plann. Manag.

    (2010)
  • T. Chai et al.

    Root mean square error (Rmse) or mean absolute error (Mae)?–Arguments against avoiding Rmse in the literature

    Geosci. Model Dev. (GMD)

    (2014)
  • H. De Bruin et al.

    The Priestley-Taylor evaporation model applied to a large, shallow lake in The Netherlands

    J. Appl. Meteorol.

    (1979)
  • S. Deswal et al.

    Artificial neural network based modeling of evaporation losses in reservoirs

    Int. J. Math. Phys. Eng. Sci.

    (2008)
  • E. Dogan et al.

    Evaporation amount estimation using Genetic algorithm and Neural networks

    Proceedings of 5th International Symposium on Intellegent Manufacturing Systems

    (2006)
  • M.E.-S. El-Mahdy et al.

    Development of mass-transfer evaporation model for Lake Nasser, Egypt

    Journal of Water and Climate Change

    (2019)
  • M.E. El-Mahdy
    (2011)
  • M.E. El-Mahdy

    Recent Changes of Lake Nasser Hydrology

    (2014)
  • M.E. El-Mahdy et al.

    Investigating the Water Quality of the Water Resources Bank of Egypt: Lake Nasser. Conventional Water Resources And Agriculture In Egypt

    (2018)
  • M. Elsawwaf et al.

    Analysis of the climate variability on Lake Nasser evaporation based on the Bowen ratio energy budget method

    J. Environ. Biol.

    (2012)
  • B. Entz

    Lake Nasser and Lake Nubia. The Nile, Biology of an Ancient

    (1976)
  • A.R. Erler et al.

    Simulating climate change impacts on surface water resources within a lake‐affected region using regional climate projections

    Water Resour. Res.

    (2019)
  • Cited by (11)

    • Projection of temperature and precipitation in the Mediterranean region through multi-model ensemble from CMIP6

      2022, Atmospheric Research
      Citation Excerpt :

      ANN models try to match the non-linear relationship between parameters to a complex problem. Although there are different architectures to create different ANN models, the feed-forward error backpropagation artificial neural network (FFEBPANN) technique is one of the most widely used architectures (El-Mahdy et al., 2021). Therefore, the FFEBPANN technique is used in this study.

    • A Unified index of water resources systems vulnerability assessment – Translating the theoretical approach into a simple tool to assess climate change Impact: Case study in Limpopo River Basin, Africa: A Unified index of water resources systems vulnerability assessment

      2022, Ain Shams Engineering Journal
      Citation Excerpt :

      The proposed method used a collection of indicators to measure the structural, functional, and gross importance of each node in WDN. The studies of the vulnerability of water resources concerning climate change were performed on many areas with different circumstances [27–38]. Keyhanpour, Musavi Jahromi [39] simulated sustainable management of WRS to find the influence of socio-economic development on energy, food, and water resources in the state of Khuzestan, Iran.

    • Impact of climate change on rainfall variability in the Blue Nile basin

      2022, Alexandria Engineering Journal
      Citation Excerpt :

      But floods are striking other places [1]. Recently become apparent that both drought and flood have increased in both frequency and size [2]. Funk, Senay [3] expected a decrease in the rainfall in addition to higher irregularity in the BNB in the future.

    View all citing articles on Scopus
    View full text