Using radar data to extend the lead time of neural network forecasting on the river Ping
Neural networks (NNs) and other data-driven methods are appearing with increasing frequency in the literature for the prediction of river levels or flows. Many of these data-driven models are tested on short lead times where they perform very well. There have been much fewer documented attempts at p...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Journal |
| Published: |
2018
|
| Subjects: | |
| Online Access: | https://www.scopus.com/inward/record.uri?partnerID=HzOxMe3b&scp=77955373000&origin=inward http://cmuir.cmu.ac.th/jspui/handle/6653943832/50745 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Chiang Mai University |
| Summary: | Neural networks (NNs) and other data-driven methods are appearing with increasing frequency in the literature for the prediction of river levels or flows. Many of these data-driven models are tested on short lead times where they perform very well. There have been much fewer documented attempts at predicting floods at longer, more useful lead times from a flood warning and civil protection perspective. In this paper NN flood forecasting models for the Upper Ping catchment at Chiang Mai are developed. Simple input determination methods are used to automate the process of which inputs to select for inclusion in the model. Lead times of 6, 12 and 18 hours are tested. Radar data inputs are then added to these NN models to see whether the lead time of the prediction can be increased. The models without radar data show reasonable forecasting ability up to 18 hours ahead but the addition of radar extends the lead times up to 36 hours ahead for the prediction of the rising limb of the hydrograph and the flood peak. |
|---|