|
1.INTRODUCTIONA number of coastal areas in the world are threaten by the increasingly environmental problems such as storm surge and sea level rise1. The effectively approaches of intervention are required to provide sustainable protection against the risk of flooding1, 2. Sand nourishment belongs to a kind of soft approaches which is to place the sand imported for strengthening the beach1,3. Sand nourishment is considered as a more environmentally friendly, economical, and efficient counter measure against sea flooding. Moreover, an additional appealing area can be created by the nourishment for the coast tourism4. So sand nourishments are widely applied to protect coastal areas5,6. With beach nourishment increasing worldwide3, the morphological evolution prediction is of major importance7,8. However, previous studies are limited due to the lack of the terrain observations of sand nourishment3,7,9. The nourishment morphology prediction output the future data by inputting the historical data, which is actually a sequence prediction problem. The deep learning methods are capable of considering spatiotemporal changes from historical data and obtain the predicted data by the time-series monitoring data. The prediction model can be gradually trained and adjusted to reduce prediction error and achieve optimal performance10,11. The deep learning has many applications such as predicting sea surface temperature, sea level changes, precipitation distribution, air quality and et al.10,12-15. To make use of the temporal and spatial information from the survey data, a data-driven method combining the long short term memory (LSTM) network and convolutional neural network (CNN) is proposed to predict the nourishment morphology evolution. Section 2 describes the survey data and the model combining LSTM and CNN. Section 3 describes and analyses the experimental data and the results. Finally, the conclusion is drawn in Section 4. 2.DATA AND METHODOLOGY2.1Study dataThe mega-nourishment known as the Sand Engine comprising 21.5 million m3 sand nourishment in the Netherlands was constructed in 2011 for coastal protection and coastline maintenance as well as for ecology and recreational purposes2,4. The nourishment terrain is observed monthly since summer 20112. The surveyed area was 4.7 by 1.6 km as shown in Figure 1a. The observed data are realigned to shoreline orientation as Figure 1b. As can be seen in Figure 1c, the collected x, y, z point data are transformed to a 20 by 20 m grid through the Inverse Distance Weighting (IDW) interpolation. 2.2Nourishment morphology prediction by combining CNN and LSTMThe spatial correlation of nourishment terrain is acquired by the CNN model. The deep neural layers are capable of being used to obtain the spatial correlation features. This architecture makes CNN perform well for obtaining features because it has the strengths of local connectivity and spatial arrangement between layer nodes. The CNN mainly consists of two kinds of layers. The first is the convolutional layers, which is used for obtaining the spatial feature. The second is the pooling layers, which is employed to reduce the number of layer links and improve computing efficiency. The spatial features extracted by the above network are put into the LSTM model after flattening to one-dimensional vector. The LSTM network has the structure of memory units, which is capable of storing long term dependencies. The memory unit contains three kinds of gates. The input gate can be used to determine what new features to store. The output gate can control how to deliver output information. The forgotten gate is adopted for reducing some features. 2.3Predicting nourishment terrainThe nourishment prediction result is obtained by the fully connected layer following the above CNN and LSTM. The prediction framework is shown as Figure 2. The convolutional kernel size is 3×3 with stride 1×1. ‘‘Same-Padding’’ is performed during the convolution processes. The activation function of ReLU is used for convolutional layers and LSTM layer. The stateful LSTM has the ability of maintaining the information through long time series16. The time lag is set to three. Mean squared error (MSE) is employed as loss function during the training process. 3.RESULTSThe nourishment observations from August 2011 to September 2015 are used in this research. Twenty-three data from August 2011 to April 2014 is used to train the model. The remaining nine data in July 2014 to September 2015 is used to test the model. Each three past grid data is used as one input sample. A predicted next terrain grid is the output. A one-step prediction model based on the observations is created. A total of 30 epochs is run for training model. The recursive method is used for multiple-step prediction. Figure 3 shows the root of mean squared error (RMSE) variation of each model over prediction time. And the proposed model results are compared with four classic models, namely, CNN, RNN, LSTM, and the BP neural network. The same dataset and the recursive prediction process are adopted by each model as the proposed model. The average RMSE of the five models are shown as figure 4. The best performance is achieved by the proposed model in this paper as shown in Figures 3 and 4. Figure 5-7 show the observations and predictions of this proposed model at three different dates for comparison. The measurements and the predicted terrain are generally consistent. The differences mainly exist at steep slope. 4.CONCLUSIONSand nourishment is widely used for protecting coastal areas against the flooding risk. The nourishment prediction is important because of increasingly nourishments worldwide. To make use of temporal and spatial information, we carry out nourishment prediction by combining LSTM and CNN. The spatial features of nourishment are obtained by CNN from the observation grids. The time dependency is tackled by LSTM. The nourishment prediction result is obtained by the fully connected layer following the above CNN and LSTM. The results show that the best performance is achieved by the proposed model in this paper compared with other methods. ACKNOWLEDGMENTSThis research has been supported by the National Natural Science Foundation of China (Grant No. 41977394); the Major Project of Science and Technology of Yunnan Province (Grant No. 202002AE090010). REFERENCESBrière, C., Janssen, S. K., Oost, A. P., Taal, M. and Tonnon, P. K.,
“Usability of the climate-resilient nature-based sand motor pilot, The Netherlands,”
Journal of coastal conservation, 22 491
–502
(2018). https://doi.org/10.1007/s11852-017-0527-3 Google Scholar
Stive, M. J., de Schipper, M. A., Luijendijk, A. P., Aarninkhof, S. G., van Gelder-Maas, C., van Thiel de Vries, J. S., de Vries, S., Henriquez, M., Marx, S. and Ranasinghe, R.,
“A new alternative to saving our beaches from sea-level rise: The sand engine,”
Journal of Coastal Research, 29 1001
–8
(2013). https://doi.org/10.2112/JCOASTRES-D-13-00070.1 Google Scholar
Ludka, B., Guza, R. and O’Reilly, W.,
“Nourishment evolution and impacts at four southern California beaches: A sand volume analysis,”
Coastal Engineering, 136 96
–105
(2018). https://doi.org/10.1016/j.coastaleng.2018.02.003 Google Scholar
Taal, M., Loffler, M., Vertegaal, C., Wijsman, J., Van der Valk, L. and Tonnon, P.,
“[Development of Sand Motor: Concise Report Describing the First Four Years of the Monitoring and Evaluation Programme],”
Deltares, Delft,
(2016). Google Scholar
de Schipper, M. A., de Vries, S., Ruessink, G., de Zeeuw, R. C., Rutten, J., van Gelder-Maas, C. and Stive, M. J.,
“Initial spreading of a mega feeder nourishment: Observations of the Sand Engine pilot project,”
Coastal Engineering, 111 23
–38
(2016). https://doi.org/10.1016/j.coastaleng.2015.10.011 Google Scholar
Mulder, J. P. and Tonnon, P. K.,
““Sand Engine”: Background and design of a mega-nourishment pilot in the Netherlands,”
in Coastal Engineering Proc.,
35
(2011). Google Scholar
Luijendijk, A. P., Ranasinghe, R., de Schipper, M. A., Huisman, B. A., Swinkels, C. M., Walstra, D. J. and Stive, M. J.,
“The initial morphological response of the Sand Engine: A process-based modelling study,”
Coastal Engineering, 119 1
–14
(2017). https://doi.org/10.1016/j.coastaleng.2016.09.005 Google Scholar
Davidson, M. A., Turner, I. L., Splinter, K. D. and Harley, M. D.,
“Annual prediction of shoreline erosion and subsequent recovery,”
Coastal Engineering, 130 14
–25
(2017). https://doi.org/10.1016/j.coastaleng.2017.09.008 Google Scholar
Arens, S. M., Mulder, J. P., Slings, Q. L., Geelen, L. H. and Damsma, P.,
“Dynamic dune management, integrating objectives of nature development and coastal safety: Examples from the Netherlands,”
Geomorphology, 199 205
–13
(2013). https://doi.org/10.1016/j.geomorph.2012.10.034 Google Scholar
Qin, D., Yu, J., Zou, G., Yong, R., Zhao, Q. and Zhang, B.,
“A novel combined prediction scheme based on CNN and LSTM for urban PM 2.5 concentration,”
IEEE Access, 7 20050
–9
(2019). https://doi.org/10.1109/Access.6287639 Google Scholar
Sejnowski, T. J.,
“[The Deep Learning Revolution],”
The MIT Press, Cambridge
(2018). https://doi.org/10.7551/mitpress/11474.001.0001 Google Scholar
Wen, C., Liu, S., Yao, X., Peng, L., Li, X., Hu, Y. and Chi, T.,
“A novel spatiotemporal convolutional long short-term neural network for air pollution prediction,”
Science of the Total Environment, 654 1091
–9
(2019). https://doi.org/10.1016/j.scitotenv.2018.11.086 Google Scholar
Ma, C., Li, S., Wang, A., Yang, J. and Chen, G.,
“Altimeter observation-based eddy nowcasting using an improved Conv-LSTM network,”
Remote Sensing, 11 783
–800
(2019). https://doi.org/10.3390/rs11070783 Google Scholar
Yang, Y., Dong, J., Sun, X., Lima, E., Mu, Q. and Wang, X.,
“A CFCC-LSTM model for sea surface temperature prediction,”
IEEE Geoscience and Remote Sensing Letters, 15 207
–11
(2017). https://doi.org/10.1109/LGRS.2017.2780843 Google Scholar
Xingjian, S., Chen, Z., Wang, H., Yeung, D. Y., Wong, W. K. and Woo, W. C.,
“Convolutional LSTM network: A machine learning approach for precipitation nowcasting,”
Advances in Neural Information Processing Systems, 802 10
(2015). Google Scholar
Gupta, S. and Dinesh, D. A.,
“Resource usage prediction of cloud workloads using deep bidirectional long short term memory networks,”
in IEEE International Conference on Advanced Networks and Telecommunications Systems (ANTS),
1
–6
(2017). Google Scholar
|