Voorspellingen Ripcurrent forecasting
Ripcurrent forecasting
| 
| 
Since 2005 the editor(s) of this website did several efforts to develope a method or model to predict when rips are going to be dangerous. In this article we
look back and try to evaluate this proces. This article is reviewing the efforts of www.muien.nl only.
The reason why these efforts were done is that quite soon after this website was launched it became clear that forecasting dangerous situations could inform the public and beach visitors in time. The idea was that well informed visitors should become aware and will not or less do stupid things. However predicting is one thing, increasing the public awareness is another. A couple of years ago the national radio station gave a warning that the rips should be dangerous that day. However, this message was based on a situation of many rescues at a permanent rip the day before. Nothing happened that day and the experienced beach visitors and locals did lift their shoulders. If these warnings happen more nobody will listen to it and the effect is zero.Later on I will put down some requirements for a rip warning system for the Dutch coast based on my personal opinion.
What efforts were done and what results were obtained:
| Year | Experiment www.muien.nl | Conclusions |
| 2005 | An experiment was done in coöporation with the lifeguard and Wl-Delft. Argus camera's were used to determine ripcurrent. Lifeguard did log the wave circumstances and their opinion of the severety of the ripcurrents. However the Argus data was not available and the amount of observations was too low for reliable conclusions. | The feeling did arise that at wave heights of above 60-90 cm the ripcurrents became risky. |
| 2007 | In the background the LURCS RIP RISK FACTOR was set-up for determination. The factors were evaluated against the warning signals of the lifeguard. | A good match was possible on a day base. The available data for wave height and direction was difficult to find. |
| 2008 | A further attempt was done with the LURCS RIP RISK FACTOR. A forecast for the day was developed and distributed to the lifeguard as a kind of beach message. | This method was not reliable enough and very time consuming. |
| 2009 | A change of course: a virtual model of a rip was made in a spreadsheet. Wave and other data was obtained from buoys in front of IJmuiden harbour and Petten-South. In a fixed rip in front of the lifeguard station many rescues took place. This model which calculates the set-up on a sand-bar and in the rip determines a rip current based on the difference of these 2 set-ups. The wave characteristics are assumed to be lineair. The prediction was made afterwards. The experiment was held in the background. | A very good evaluation with the actual moments of rescues showed a very good result. Over 90% of the rescues took place in the moments the models predicted ripcurrents. Disadvantage is that the height of the bar and deepness of the rip had to be guessed. Regurlary calibration of the model was necessary. See publication Muistroom 5 (In Dutch) |
| 2010 | The model of 2009 was extended, more virtual rips with different sand-bars and rips were coupled to a buy data acquisition system of Rijkswaterstaat. (MFPS). In this model wave forecast was used also. This experiment took place in the background. Evaluation was done by personal observation of the rips. | During the summer the model was optimised for all kind of sea conditions. It appeared that the model was reliable enough for publishing it on a real-time base. |
| 2011 | This summer the model was launched on internet. Only during certain hours a prediction was given by means of graphs. The prediction was based on a average ripcurrent in % of 4 rips in the 1st bar and 2 rips in the beachbar. During the summer the model was changed to all rips in the 1st bar. During the last week of August the model was used for predicting the deployment moments of the ripcurrent survey of Deltares at Egmond aan Zee. | Prelimenary conclusion of ths summer is that the rips in Egmond were very dynamic or did not appear at all. Main reason for this was the effects of the sand suppletion in front of Egmond. During the different weather condition the wave types were studied and compared with the wave factor of the model. The used shape for the sandbar appeared to be too simple. (linear slope). Also the tidal upset was changing a lot and it proved that a refresh time of 10 minutes was needed. Problematic is the estimatin of the wave height leaving bar 2 at low tide. During the deployments at the end of August the rips were active, however the circumstances were not so dangerous. The amount of current the model predicts is about a factor 4 too high. During the summer the yellow flag was hoisted 34 times and the red flag 12 times. After the information of the lifeguard about rescues is received a short evaluation will be published on this site. |
| 2012 | The model will be redesigned. The slope will be made as is natural, more wave data will be used, momentarily and predictions of 1,2,3 and 4 hours in advance will be generated. A friendlier way of presenting the data will be developed. | This model will probably be vissible on the internet. More news will be launched at spring 2012. |
