Flood forecasting on the Humber River using an artificial neutral network approach
Bibliographic record
Abstract
In order to provide flood warnings to the residents living along the various sections of the Humber River Basin, the Water Resources Management Division (WRMD) of Department of Environment and Conservation, Government of Newfoundland and Labrador has generated flow forecasts for this basin over the years by means of several rainfall-runoff models. The first model used is the well-known Streamflow Synthesis and Reservoir Regulation Model (SSARR) which is a deterministic model that accounts for some or all of the hydrologic factors responsible for runoff in the basin. However, the accuracy of the model became worse over the years. Although it was calibrated well in the beginning, recalibration of the model has not been very successful. In addition, the model cannot take into account the snowmelt effect from the Upper Humber basin. The next model is the Dynamic Regression model, a statistically based model that uses the time series of historic flows and climate data of the basin to generate a forecast. This model was tried during the late 1990s to early 2000s. This model was found to provide better forecasts than the SSARR model, but it also does not take into account the snowmelt effect from the upper regions of the Humber River. The third model tried by the WRMD was an in-house Routing model. This method uses a series of water balance equations which can be easily implemented on a spread sheet at each gauging station. However, calibration is done subjectively and the forecast obtained for the snowy region of the Upper Humber is still a problem. In view of the foregoing issues with the above models, a better model that is easy to use and calibrate, provides accurate forecasts, and one that can take into account the snowmelt effects is required. Since 2008, the WRMD has been using the statistically based Dynamic Regression Model on an interim basis until a replacement model could be developed. -- This thesis presents the development of artificial neural network (ANN) models for river flow forecasting for the Humber River Basin. Two types of ANN were considered, general regression neural network (GRNN) and the back propagation neural network (BPNN). GRNN is a nonparametric method with no training parameters to be adjusted during the training process. BPNN on the other hand has several parameters such as the learning rate, momentum, and calibration interval, which can be adjusted during the training to improve the model. A design of experiment (DOE) approach is used to study the effects of the various inputs and network parameters at various stages of the network development to obtain an optimal model. One day ahead forecasts were obtained from the two ANNs using air temperature, precipitation, cumulative degree-days, and flow data all suitably lagged (i.e. of 1 day or 2 day before) as inputs. It was found that the GRNN model produced slightly better forecasts than the BPNN for the Upper Humber and both models performed equally well for the Lower Humber. The ANN approach also produced much better forecasts than the routing model developed by the WRMD but was not much better than the dynamic regression model except for the Upper Humber.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".