• Journal of Korea Water Resources Association
• /
• v.56 no.11
• /
• pp.775-784
• /
• 2023

With the increasing trend of extreme rainfall that exceeds the design frequency of man-made structures due to extreme weather, it is necessary to review the safety of agricultural reservoirs designed in the past. However, there are no local government-managed reservoirs (13,685) that can be discharged in an emergency, except for reservoirs over a certain size under the jurisdiction of the Korea Rural Affairs Corporation. In this case, it is important to quickly deploy a mobile siphon to the site for preliminary discharge, and this study evaluated the applicability of a mobile siphon with a diameter of 200 mm, a minimum water level difference of 6 m, 420 (m²/h), and 10,000 (m²/day), which can perform both preliminary and emergency discharge functions, to the Yugum Reservoir in Gyeongju City. The test bed, Yugum Reservoir, is a facility that was completed in 1945 and has been in use for about 78 years. According to the hydrological stability analysis, the lowest height of the current dam crest section is 27.15 (EL.m), which is 0.29m lower than the reviewed flood level of 27.44 (EL.m), indicating that there is a possibility of lunar flow through the embankment, and the headroom is insufficient by 1.72 m, so it was reviewed as not securing hydrological safety. The water level-volume curve was arbitrarily derived because it was difficult to clearly establish the water level-flow relationship curve of the reservoir since the water level-flow measurement was not carried out regularly, and based on the derived curve, the algorithm for operating small and medium-sized old reservoirs was developed to consider the pre-discharge time, the amount of spillway discharge, and to predict the reservoir lunar flow time according to the flood volume by frequency, thereby securing evacuation time in advance and reducing the risk of collapse. Based on one row of 200 mm diameter mobile siphons, the optimal pre-discharge time to secure evacuation time (about 1 hour) while maintaining 80% of the upper limit water level (about 30,000 m²) during a 30-year flood was analyzed to be 12 hours earlier. If the pre-discharge technology utilizing siphons for small and medium-sized old reservoirs and the algorithm for reservoir operation are implemented in advance in case of abnormal weather and the decision-making of managers is supported, it is possible to secure the safety of residents in the risk area of reservoir collapse, resolve the anxiety of residents through the establishment of a support system for evacuating residents, and reduce risk factors by providing risk avoidance measures in the event of a reservoir risk situation.

• Journal of Korea Water Resources Association
• /
• v.57 no.7
• /
• pp.437-449
• /
• 2024

Rainfall-runoff prediction studies using deep learning while considering catchment attributes have been gaining attention. In this study, we selected two models: the Transformer model, which is suitable for large-scale data training through the self-attention mechanism, and the LSTM-based multi-state-vector sequence-to-sequence (LSTM-MSV-S2S) model with an encoder-decoder structure. These models were constructed to incorporate catchment attributes and predict the inflow of 10 multi-purpose dam watersheds in South Korea. The experimental design consisted of three training methods: Single-basin Training (ST), Pretraining (PT), and Pretraining-Finetuning (PT-FT). The input data for the models included 10 selected watershed attributes along with meteorological data. The inflow prediction performance was compared based on the training methods. The results showed that the Transformer model outperformed the LSTM-MSV-S2S model when using the PT and PT-FT methods, with the PT-FT method yielding the highest performance. The LSTM-MSV-S2S model showed better performance than the Transformer when using the ST method; however, it showed lower performance when using the PT and PT-FT methods. Additionally, the embedding layer activation vectors and raw catchment attributes were used to cluster watersheds and analyze whether the models learned the similarities between them. The Transformer model demonstrated improved performance among watersheds with similar activation vectors, proving that utilizing information from other pre-trained watersheds enhances the prediction performance. This study compared the suitable models and training methods for each multi-purpose dam and highlighted the necessity of constructing deep learning models using PT and PT-FT methods for domestic watersheds. Furthermore, the results confirmed that the Transformer model outperforms the LSTM-MSV-S2S model when applying PT and PT-FT methods.

• Journal of The Korean Society of Grassland and Forage Science
• /
• v.5 no.1
• /
• pp.22-32
• /
• 1985

Optimum pasture management during the summer season is an important factor to maintain good regrowth and persistence of pasture in Korea. This experiment was carried out to investigate the effects of the cutting management on the dead plant, weed appearance, regrowth and carbohydrate reserves in stubble, and dry matter yield of tall fescue dominated pasture during the mid summer season. For the test, a split plot design with 4 replications was treated with 2 different the third cutting times (July 12 and Aug. 4) as the mainplots, and 3 different cutting heights (3, 6 and 9 cm) at the third cut as the subplots, and the experiment was done at the experimental field of the Livestock Experiment Station, in Suweon, 1984. The results obtained are summarized as follows: 1. Considering the meteorological conditions during the experimental period, the temperature was a little higher by $2^{\circ}C$ than that of average year, especially the first and second decade of August were high. And the precipitation of 1984 tended to be low when compared with the average year. 2. Temperature of soil surface and underground tended to increase by $1-3^{\circ}C$ as the stubble height was low during the summer season. 3. Regrowth leaf length and leaf area after the third cut increased significantly with the high cutting height at the third cut. 4. A significant higher total nonstructural carbohydrate (TNC) content in stubble after the third cut was observed in the high stubble cut on July 12. The results indicate that the high stubble height reserves more carbohydrates for early regrowth stage after the third cut when compared with the low stubble. On Aug. 4, however, the recovery of TNC contents after the third cut was not effective due to high temperature and rainfall. 5. The percentage of dead plant after the third cut was found to be high with the low cutting height during the mid summer season (p<0.05). 6. With the low stubble height on July 12 cut, it was appeared that the percentage of weed was significantly increased (p<0.05), and main weeds appeared after the third cut were Echinochloa crusgalli>Digitaria sanguinalis>Cyperus iria>Rumex crispus, and so on. In case of cut on Aug. 4, weed appearance was no difference at three cutting heights. 7. Dry matter yield at the third cut was increased in the plot of cutting on Aug. 4 and stubble height (p<0.05). However, yields at the fourth and fifth cut were increased with high stubble height (p<0.05), regardless of harvest time. 8. In total dry matter yield after the third cut, there was no significant difference between the cutting time and forage yield. However, total yield on July 12 was increased with the high stubble height (p<0.05). 9. From the above results, it is suggested that the 9 cm cutting height during the mid summer season is the most effective for good regrowth, weed control and forage yield of tall fescue dominated pasture.