1박사과정생, 충남대학교 토목공학과, 대전광역시 유성구 대학로 99, 34134, 대한민국
2교수, 충남대학교 토목공학과, 대전광역시 유성구 대학로 99, 34134, 대한민국
1Ph.D Candidate, Department of Civil Engineering, Chungnam National University, 99 Daehak-ro, Yoosung-gu, 34134 Daejeon, South Korea
2Professor, Department of Civil Engineering, Chungnam National University, 99 Daehak-ro, Yoosung-gu, 34134 Daejeon, South Korea
Copyright © 2023 GeoAI Data Society
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Conflict of Interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
Funding Information
None.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
Value | |
---|---|
Training | Day 1 to Day 10,000 |
Validation | Day 10,001 to Day 15,000 |
Test (prediction) | Day 15,001 to Day 17,532 |
Sort | Filed | Subcategory#1 | Subcategory#2 |
---|---|---|---|
Essential | *Title | Dam inflow of Soyangdam | |
*DOI name | https://doi.org/10.22761/GD.2023.0016 | ||
*Category | Hydrology | ||
Abstract | |||
*Temporal Coverage | 1974 January 1 to 2021 December 31 | ||
*Spatial Coverage | Address | Chuncheon Kangwondo Korea Rep. | |
WGS84 Coordinates | 37.945556, 127.814444 | ||
*Personnel | Name | YS. JO | |
Affiliation | K-water | ||
jyslord@kwater.or.kr | |||
*CC License | None | ||
Optional | *Project | None | |
*Instrument | None |
Data group | Data set |
CASE scenario |
CASE #N0 | |||
---|---|---|---|---|---|---|
Independent variable | Independent | Dependent | ||||
Rainfall (R) | Basic data | R(t) | R(t) | Q(t) | #1 | |
Time lagged data | R(t-1) | R(t)-R(t-1) | Q(t) | #2 | ||
R(t-2) | R(t)-R(t-2) | Q(t) | #3 | |||
R(t-3) | R(t)-R(t-3) | Q(t) | #4 | |||
R(t-5) | R(t)-R(t-5) | Q(t) | #5 | |||
R(t-10) | R(t)-R(t-10) | Q(t) | #6 | |||
R(t-30) | R(t)-R(t-30) | Q(t) | #7 | |||
Moving averaged data | R(MA2) | R(t)-R(t-3), R(MA2) | Q(t) | #8 | ||
R(MA3) | R(t)-R(t-3), R(MA3) | Q(t) | #9 | |||
R(MA5) | R(t)-R(t-3), R(MA5) | Q(t) | #10 | |||
R(MA10) | R(t)-R(t-3), R(MA10) | Q(t) | #11 | |||
R(t)-R(t-10), R(MA10) | Q(t) | #12 | ||||
R(t)-R(t-10), R(MA2, 3, 5, 10) | Q(t) | #13 | ||||
R(t)-R(t-30), R(MA2, 3, 5, 10) | Q(t) | #14 | ||||
Rainfall (R) and inflow (Q) | Time lagged data | Q(t-1) | R(t), Q(t-1) | Q(t) | #15 | |
Q(t-3) | R(t)-R(t-3), Q(t-1) | Q(t) | #16 | |||
Q(t-5) | R(t)-R(t-10), Q(t-1) | Q(t) | #17 | |||
Q(t-30) | R(t)-R(t-3), Q(t-1)-Q(t-5) | Q(t) | #18 | |||
R(t)-R(t-5), Q(t-1)-Q(t-5) | Q(t) | #19 | ||||
R(t)-R(t-10), Q(t-1)-Q(t-10) | Q(t) | #20 | ||||
R(t)-R(t-30), Q(t-1)-Q(t-30) | Q(t) | #21 | ||||
Differenced delta value | delQ [Q(t-n)-Q(t-(n-1))] | R(t)-R(t-3), Q(t-1), delQ(t-1) | Q(t) | #22 | ||
Component separation | Bf(t-1) | Sf(t-1) | R(t), Bf(t-1) | Q(t) | #23 | |
Bf(t-2) | Sf(t-2) | R(t), Bf(t-1), Sf(t-1) | Q(t) | #24 | ||
- Baseflow (Bf) | Bf(t-3) | Sf(t-3) | R(t)-R(t-3), Bf(t-1), Sf(t-1) | Q(t) | #25 | |
- Surfaceflow (Sf) | Bf(t-5) | Sf(t-5) | R(t)-R(t-5), Bf(t-1), Sf(t-1) | Q(t) | #26 | |
R(t)-R(t-5), Bf(t-1), Sf(t-1-t-5) | Q(t) | #27 | ||||
Rainfall (R), inflow (Q) and meteorological data | Temperature, evaporation | Tavg(t) (averaged temperature) | R(t), Tavg(t) | Q(t) | #28 | |
R(t), Evap(t) | Q(t) | #29 | ||||
Evap(t) | R(t), Tavg(t), Evap(t) | Q(t) | #30 | |||
(evaporation) | R(t)-R(t-3), Tavg(t), Evap(t) | Q(t) | #31 | |||
R(t)-R(t-5), Tavg(t), Evap(t) | Q(t) | #32 | ||||
R(t)-R(t-5), Q(t-1), Tavg(t), Evap(t) | Q(t) | #33 | ||||
R(t)-R(t-10), Tavg(t), Evap(t) | Q(t) | #34 | ||||
R(t-t-3), B(t-1), Sf(t-1), Evap(t), Tavg(t) | Q(t) | #35 | ||||
R(t)-R(t-10), Bf(t-1), Sf(t-1-t-5), Evap(t), Tavg(t) | Q(t) | #36 |
Model | Hyperparameter |
---|---|
SVR | Kernel type (rbf), gamma (0.02), C-value (1300), epsilon (1), degree (-) |
RF | Bootstrap (true), max_deapth (5), max_feature (auto), n_estimator (100) |
LightGBM | Colsamples_bytree (0.8), min_child_samples (20), max-depth (3), n_estimators (100), num_leaves (20) |
LSTM | Layer (2), nodes (512), activation function (1-tanh, 2-relu), optimizer (Adam), batch size (512) |
TCN | Layer (3), filters (512-512-512)), kernel size (4,4,4), activation function (1-relu, 2-relu, 3-relu), batch size (512) |
Content | Value |
---|---|
River basin | Bukhan River |
Basin area | 2,703 km2 |
Length/height | 530 m/123 m |
Dam type | E.C.R.D |
Generation capacity | 200 MW |
Annual generation | 353 GWh |
Total storage volume | 2,900 million m3 |
Storage area | 70 km2 |
Annual averaged inflow (1974 to 2021) | 67.7 m3/s |
2,153 million m3 | |
Annual averaged rainfall (1974 to 2021) | 1,214 mm |
Planned basic supply | 1,468.4 million m3 |
Value | |
---|---|
Training | Day 1 to Day 10,000 |
Validation | Day 10,001 to Day 15,000 |
Test (prediction) | Day 15,001 to Day 17,532 |
No. | Data case | LR | Lasso | Ridge | SVR | RF | LGBM | XGB | LSTM | TCN | TCN-LSTM |
---|---|---|---|---|---|---|---|---|---|---|---|
#1 | R(t) | 0.404 | 0.404 | 0.404 | 0.463 | 0.448 | 0.472 | 0.421 | 0.478 | 0.495 | 0.498 |
#2 | R(t~t-1) | 0.629 | 0.63 | 0.629 | 0.783 | 0.743 | 0.773 | 0.725 | 0.759 | 0.767 | 0.772 |
#3 | R(t~t-2) | 0.659 | 0.659 | 0.659 | 0.804 | 0.746 | 0.78 | 0.745 | 0.779 | 0.807 | 0.806 |
#4 | R(t~t-3) | 0.671 | 0.671 | 0.671 | 0.82 | 0.786 | 0.812 | 0.791 | 0.828 | 0.833 | 0.822 |
#5 | R(t~t-5) | 0.681 | 0.681 | 0.681 | 0.832 | 0.818 | 0.837 | 0.817 | 0.817 | 0.858 | 0.833 |
#6 | R(t~t-10) | 0.684 | 0.684 | 0.684 | 0.82 | 0.842 | 0.845 | 0.82 | 0.857 | 0.864 | 0.834 |
#7 | R(t~t-30) | 0.685 | 0.685 | 0.685 | 0.754 | 0.846 | 0.855 | 0.836 | 0.839 | 0.867 | 0.858 |
#8 | R(t~t-3), R(MA2) | 0.671 | 0.671 | 0.671 | 0.823 | 0.802 | 0.816 | 0.804 | 0.81 | 0.839 | 0.827 |
#9 | R(t~t-3), R(MA3) | 0.671 | 0.671 | 0.671 | 0.819 | 0.785 | 0.797 | 0.792 | 0.825 | 0.83 | 0.821 |
#10 | R(t~t-3), R(MA5) | 0.681 | 0.681 | 0.681 | 0.834 | 0.833 | 0.819 | 0.818 | 0.822 | 0.847 | 0.829 |
#11 | R(t~t-3), R(MA10) | 0.682 | 0.682 | 0.682 | 0.845 | 0.84 | 0.832 | 0.821 | 0.823 | 0.853 | 0.837 |
#12 | R(t~t-10), R(MA10) | 0.684 | 0.684 | 0.684 | 0.817 | 0.848 | 0.844 | 0.828 | 0.841 | 0.86 | 0.832 |
#13 | R(t~t-10), R(MA2, 3, 5, 10) | 0.682 | 0.684 | 0.684 | 0.809 | 0.834 | 0.845 | 0.833 | 0.853 | 0.865 | 0.837 |
#14 | R(t~t-30), R(MA2, 3, 5, 10) | 0.681 | 0.685 | 0.685 | 0.729 | 0.83 | 0.847 | 0.85 | Fail | 0.874 | Fail |
#15 | R(t), Q(t-1) | 0.663 | 0.663 | 0.663 | 0.783 | 0.756 | 0.759 | 0.747 | 0.793 | 0.796 | 0.79 |
#16 | R(t~t-3), Q(t-1) | 0.705 | 0.705 | 0.705 | 0.865 | 0.847 | 0.848 | 0.841 | 0.828 | 0.865 | 0.871 |
#17 | R(t~t-10), Q(t-1) | 0.708 | 0.708 | 0.708 | 0.825 | 0.863 | 0.86 | 0.829 | 0.841 | 0.866 | 0.846 |
#18 | R(t~t-3), Q(t~t-5) | 0.705 | 0.706 | 0.705 | 0.873 | 0.874 | 0.875 | 0.848 | 0.81 | 0.884 | 0.886 |
#19 | R(t~t-5), Q(t~t-5) | 0.704 | 0.704 | 0.704 | 0.846 | 0.874 | 0.878 | 0.848 | 0.85 | 0.887 | 0.851 |
#20 | R(t~t-10), Q(t~t-10) | 0.705 | 0.706 | 0.705 | 0.787 | 0.875 | 0.88 | 0.863 | 0.835 | 0.868 | 0.848 |
#21 | R(t~t-30), Q(t~t-30) | 0.702 | 0.702 | 0.702 | 0.735 | 0.862 | 0.874 | 0.856 | Fail | 0.872 | Fail |
#22 | R(t~t-3), Q(t-1), delQ(t-1) | 0.705 | 0.705 | 0.705 | 0.865 | 0.85 | 0.848 | 0.841 | 0.826 | 0.871 | 0.863 |
#23 | R(t), Bf(t-1) | 0.527 | 0.527 | 0.527 | 0.661 | 0.672 | 0.696 | 0.686 | 0.654 | 0.679 | 0.679 |
#24 | R(t), Bf(t-1), Sf(t-1) | 0.666 | 0.666 | 0.666 | 0.777 | 0.78 | 0.78 | 0.759 | 0.765 | 0.81 | 0.805 |
#25 | R(t~t-3), Bf(t-1), Sf(t-1) | 0.715 | 0.715 | 0.715 | 0.898 | 0.878 | 0.879 | 0.860 | 0.829 | 0.901 | 0.905 |
#26 | R(t~t-5), Bf(t-1), Sf(t-1) | 0.714 | 0.714 | 0.714 | 0.882 | 0.894 | 0.884 | 0.865 | 0.832 | 0.887 | 0.86 |
#27 | R(t~t-5), Bf(t-1), Sf(t-1~t-5) | 0.711 | 0.711 | 0.711 | 0.842 | 0.894 | 0.889 | 0.869 | 0.851 | 0.891 | 0.869 |
#28 | R(t), Tavg(t) | 0.408 | 0.408 | 0.408 | 0.502 | 0.456 | 0.486 | 0.442 | 0.508 | 0.507 | 0.497 |
#29 | R(t), Evap(t) | 0.41 | 0.41 | 0.41 | 0.488 | 0.51 | 0.505 | 0.48 | 0.482 | 0.498 | 0.492 |
#30 | R(t), Tavg(t), Evap(t) | 0.41 | 0.41 | 0.41 | 0.507 | 0.46 | 0.507 | 0.43 | 0.516 | 0.513 | 0.498 |
#31 | R(t~t-3), Tavg(t), Evap(t) | 0.675 | 0.675 | 0.675 | 0.831 | 0.798 | 0.822 | 0.796 | 0.828 | 0.839 | 0.834 |
#32 | R(t~t-5), Tavg(t), Evap(t) | 0.687 | 0.687 | 0.687 | 0.842 | 0.824 | 0.850 | 0.824 | 0.839 | 0.843 | 0.836 |
#33 | R(t~t-5), Q(t-1), Tavg(t), Evap(t) | 0.713 | 0.713 | 0.713 | 0.866 | 0.868 | 0.859 | 0.846 | 0.844 | 0.867 | 0.841 |
#34 | R(t~t-10), Tavg(t), Evap(t) | 0.692 | 0.693 | 0.692 | 0.826 | 0.845 | 0.851 | 0.821 | 0.827 | 0.864 | 0.842 |
#35 | R(t~t-3), B(t-1), Sf(t-1), Evap(t), Tavg(t) | 0.725 | 0.726 | 0.725 | 0.905 | 0.891 | 0.884 | 0.868 | 0.839 | 0.894 | 0.900 |
#36 | R(t~t-10), Bf(t-1), Sf(t-1~t-5), Evap(t), Tavg(t) | 0.722 | 0.722 | 0.722 | 0.801 | 0.896 | 0.884 | 0.876 | 0.845 | 0.866 | 0.849 |
Maximum | 0.725 | 0.726 | 0.725 | 0.905 | 0.896 | 0.889 | 0.876 | 0.857 | 0.901 | 0.905 | |
Average | 0.654 | 0.654 | 0.654 | 0.782 | 0.791 | 0.799 | 0.778 | 0.782 | 0.812 | 0.796 | |
Minimum | 0.404 | 0.404 | 0.404 | 0.463 | 0.448 | 0.472 | 0.421 | 0.478 | 0.495 | 0.492 |
Sort | Filed | Subcategory#1 | Subcategory#2 |
---|---|---|---|
Essential | *Title | Dam inflow of Soyangdam | |
*DOI name | |||
*Category | Hydrology | ||
Abstract | |||
*Temporal Coverage | 1974 January 1 to 2021 December 31 | ||
*Spatial Coverage | Address | Chuncheon Kangwondo Korea Rep. | |
WGS84 Coordinates | 37.945556, 127.814444 | ||
*Personnel | Name | YS. JO | |
Affiliation | K-water | ||
jyslord@kwater.or.kr | |||
*CC License | None | ||
Optional | *Project | None | |
*Instrument | None |
ML, machine learning; DL, deep learning; SVR, Support Vector Regression; RF, Random Forest; LSTM, Long Short-Term Memory; TCN, Temporal Convolutional Network.
ML, machine learning; DL, deep learning; LR, Linear Regression; SVR, Support Vector Regression; RF, Random Forest; LGBM, Light Gradient Boosting Model; XGB, eXtream Gradient Boosting Model; LSTM, Long Short-Term Memory; TCN, Temporal Convolutional Network.