'To what extent does the River Lyn conform to the Bradshaw model of River characteristics?'
Introduction.Aims of investigation + my key questions.
The aim of my investigation is to see ‘To what extent does the River Lyn conform to the Bradshaw model of River characteristics?’ To help me answer this main question I have split it up into key questions. They are:
- Does the size and speed of the River increase going downstream? And therefore does the discharge increase as you go downstream?
- Does the gradient decrease as you go downstream?
- Does the load particle size decrease as you go downstream?
Background information.
The River Lyn has two channels which meet about two thirds along the River Lyn, and carry on as one until it reaches the Bristol Channel at Lynmouth. They join at Watersmeet, and from Watersmeet to the mouth the length of the River Lyn is 2.5 miles and drops 110 metres. The gradient can be up to 1 in 63 at some parts of the River Lyn. Canoeing takes place November to February (on a restricted scale), and at other times fishing is allowed.
The East and West Lyn rivers once flowed parallel to the coast eroding 660 ft deeply into the plateau, where it entered the sea at Lee Bay. With the breaching of the valley sides the East and West Lyn rivers cascaded to the shore at Lynmouth, and now there is a fossil filled Riverbed high above the active river.
On the 15th August 1952 there was a huge flood, there was high rainfall levels and all the water in the Exmoor catchment went into the River. Above Lynmouth the catchment area of a total 39.2 square miles are of gentle sloping and flat toped moors. Large boulders and rocks were carried in the flow towards Lynmouth destroying houses, roads and bridges. Many lost their lives during the flood. Changes have been made nearer the mouth of the River Lyn due to this.
The Bradshaw model of River characteristics.
This is what I will be comparing the River Lyn with to see, ‘to what extent does the River Lyn conform with the Bradshaw model of River characteristics?’
Fig. 1
Annotated sketch maps of location.
Hypothesis
Fig. 4
Method.
Methodology table
The weather, time of day and year, and the accuracy of our samples could have effected the measurements as well as the extra limitations in the table.
In addition to the equipment stated in the table, I also used a map, recording sheet, clipboard (covered for waterproofing), pencil, Wellington boots, and waterproof clothes.
What I did-how, why and the sampling techniques I used
Water + channel depth.
I measured the depth of the River, but when it got too deep I was unable to measure the depth. I will have to get secondary data to help me see how much it conforms to the Bradshaw model of River characteristics.
Fig. 5
Water + channel width.
Fig. 6
Wetted perimeter.
The way I measured ...
This is a preview of the whole essay
In addition to the equipment stated in the table, I also used a map, recording sheet, clipboard (covered for waterproofing), pencil, Wellington boots, and waterproof clothes.
What I did-how, why and the sampling techniques I used
Water + channel depth.
I measured the depth of the River, but when it got too deep I was unable to measure the depth. I will have to get secondary data to help me see how much it conforms to the Bradshaw model of River characteristics.
Fig. 5
Water + channel width.
Fig. 6
Wetted perimeter.
The way I measured the wetted perimeter, wasn’t very accurate so I drew a cross-section of the River using my other data, and then measured it from the drawing as well.
Fig. 7
Velocity.
Three measurements were taken across the River; the places were picked carefully so nothing was in the way, which could give inaccurate readings. I held the flow meter in the water and someone else timed to see how long it took for the propeller to move from one end to the other, and then an average was found. This was measured in seconds. Then to work out the proper velocity reading I used the calculation,
0.0277 + (3.2805 ÷ average time in seconds).
D90 bedload.
The D90 bedload was measuring by selecting 50 rocks, putting them in size order, and selecting the 45th largest. 100 rocks are usually selected, and then the 90th largest is recorded, but there was not enough time to do this so the 45th was used. The longest axis of the 45th largest rock was measured, and recorded. This was measured to find to what extent the load particle size of the River Lyn conformed to the Bradshaw model of River characteristics. The load particle size should increase as you go down stream.
Gradient.
The Bradshaw model of River characteristics states that the gradient should decrease as you move downstream.
Fig. 8
Cross-sectional area.
I will draw the Rivers cross section at all the places where I took the measurements, this will make sure none of the measurements are wrong, and then I can work out the Cross-sectional area. This will help work out other measurements, like the hydraulic radius.
Hydraulic radius.
Hydraulic radius is the proportion of water in a channel cross section which is in contact with the channel margin. The length of the wetted perimeter, bed length and banks in contact with the water influences it. Hydraulic radius is the measure of the efficiency of a River channel. It is worked out by the calculation cross sectional area ÷ wetted perimeter.
Discharge.
I will work out the average velocity and multiply it by the cross-sectional area. To help me see the extent the discharge conforms to the Bradshaw model of River characteristics.
The sites I took my readings at.
The results were taken at various sites along the River Lyn. The locations of these sites are in the table below.
Fig. 9
For photographs of the five sites see the data representation, analysis, and explanation.
Example of result table
This is what I filled in when I was doing my primary research. Some of the measurements had to be calculated later (like the mean velocity).
Fig. 10
Evaluating problems that occurred
Nearer the mouth of the River Lyn I was unable to take every measurement because it would have been unsafe. I had to take safety issues into consideration including the country code, paths, access availability, traffic, permission of entering certain areas, and the clothing I wore. Not all of the measurements may have been accurate, because of the weather because if it had recently been raining a lot there might be a bigger volume of water in the River. The time of day and year could affect the measurements as well. Where I took my measurements could have gave false readings, because there could have been a waterfall or meander close by, different geology of the area, and the surrounding land use. My measurements could have been inaccurate due to me inaccurately measuring them, because of the weather conditions and the amount of time I had, and the accuracy and reliability of the equipment I had. The flow and temperature of the River also could have made my results more unreliable.
Data representation, analysis and explanation.
Photographs of the five sites.
Source of the River Lyn, looking North West, down the River Lyn. 11/11/02. P.M.
Fig. 11
The first place I studied-site 1. Looking South East. 11/11/02. P.M.
Fig. 12
Site 2 , Looking North West, down the River Lyn. 11/11/02. P.M.
Fig. 13
Site 3. Looking North West, down the River Lyn. 11/11/02. P.M.
Fig. 14
Site 4, at the side of the River Lyn. 11/11/02. P.M.
Fig. 15
Site 5 nearly at the mouth of the River Lyn. 11/11/02 P.M.
Fig. 16
Calculations I used.
- See appendix for results tables *
Fig. 17
Analysing the cross sectional diagrams.
Site 1.
The banks of site one look steep. The water depth is about 6cm, and the water width is about 17cm. The cross sectional area is about 0.008m2. The wetted perimeter is 0.46m.
Site 2.
The banks in contact with the water are steep, then on the right hand side the bank gradually flattens out, whereas on the left hand side, it stays quite steep. This could be for a variety of reasons, including because there is a meander, so the outermost side would have been eroded away, there is also a path to the river, maybe from animals drinking water, which could have made the River flatter on one side then the other. The water depth is about 22cm, and the width is about 92cm. The cross sectional area is 0.1636m2. The wetted perimeter is 2.4m.
Site 3.
The banks in contact with the water are quite flat, so therefore the River is quite shallow, it is only about 21cm in depth, but about 4m62cm wide. The cross sectional area is 1.05m2 and the wetted perimeter is 9.75m.
From the data I collected from the five sites along the River Lyn, and the calculations I worked out from this data, I have produced some graphs, which are annotated.
Fig. 18
The cross sectional area can be worked out by two techniques, it can be worked out by multiplying the width (m) by the average depth (m). Or it can be worked out by drawing a cross sectional area picture drawn to scale. I calculated it from the scaled cross sectional diagram, because this is more accurate, because it shows the exact depth, rather then an average, which could be inaccurate if there is a rock on the River bed.
Fig. 19
Fig. 20
Fig. 21
Fig. 22
Fig. 23
Velocity and a statistical test.
I have chosen to do the ‘Spearman’s Rank Correlation Coefficient’ statistical test. (rs) on the relationship between the distance downstream and the velocity.
Null hypothesis = there is no relationship between the distance downstream and the velocity.
Rs can be used to discover if there is an association between two sets of measurements. The measurement levels I will use are interval levels.
Fig. 24
This scatter graph shows the distance from the source and the velocity. Now I will draw a table to work out the Spearman’s Rank Correlation Coefficient.
Fig. 25
To work out the Rs you use the calculation:
1-{(6Σd2)÷(n3-n)}
With my data the sum will be 1-{(6 x 6)÷125-5}} = 0.7
I looked up the Rs value in the significance table the published value from the significance table at the 5% level for five pairs of data is one. The calculated value is less than the published value, which allows the null hypothesis to be accepted. The data therefore suggests there is no relationship between the distance downstream and velocity.
Evaluating Spearman’s Rank Correlation Coefficient.
The graph showing velocity and distance downstream didn’t really show any linear relationship. I did the statistic test to measure the strength of the relationship. The result was 0.7. There were some problems that occurred when trying to work out the Spearman’s Rank Correlation Coefficient, including only taking five sets of readings from the River Lyn, and to get a more accurate Spearman’s Rank Correlation Coefficient then more than five sets of data were required. So if I did this again I would take more readings to get an accurate Spearman’s Rank Correlation Coefficient.
Analysis and explanation.
I have now done enough research to answer the title, ‘To what extent does the River Lyn conform to the Bradshaw model of River characteristics?’ except for the load quantity and channel bed roughness of the River Lyn. It was not possible to obtain data, because we did not have sufficient equipment or time to do this. The title has been broken down into key questions to help me answer it. I have collected data on the gradient, D90 bedload, cross sectional area, velocity, depth, width and the discharge in order to help me answer the title. For all of these I have collected the data and put it in the best possible form, e.g. diagrams, graphs, tables, photographs. I have used primary and secondary data.
The Bradshaw model of River characteristics has several parts to it, showing whether each aspect decreases or increases as you go downstream.
I chose to do a statistical test on the velocity of the River Lyn, because I wanted to know if the velocity and distance downstream had any relationship. I found out from doing a Spearman’s Rank Correlation Coefficient statistical test, that there was no relationship between the distance downstream and the velocity. The test was not very effective because I needed more readings.
I only had one anomalous result in all of my readings; this was the velocity reading for the third site. It is higher than it should be. If the anomalous result wasn’t plotted on the graph, then there would be a good positive correlation, with all points nearly on the line of best fit. This anomaly could have been for a variety of reasons, including when taking the reading with the flow metre, when shouting out when to start and stop to the person with the stopwatch, they could have had a delayed reaction time. Other reasons could be the reliability of the equipment, how rushed I was for time, or the position the flow metre was held in the water.
Fig. 26
Evaluation
and
conclusion.
My key questions and results.
My key questions are:
- Does the size and speed of the River increase going downstream? And therefore does the discharge increase as you go downstream?
- Does the gradient decrease as you go downstream?
- Does the load particle size decrease as you go downstream?
My results to these key questions:
- The photographs show the River size increases as you go downstream. From my results and graphs I can see the cross sectional area, width and mean depth also increase as you go downstream. So all of these prove that the size of the River increases as you go move from the source of the River Lyn to the mouth. From the velocity graph there is an anomalous point but there is still a positive correlation, proving that the speed of the River increases as you move downstream. So therefore the discharge also increases as you go downstream. This can also be proved by my results, and the discharge graph.
- From my results, the graph and secondary information the gradient decreases as you go downstream.
- From my results and the graph, I did on the D90 bedload, the load particle size increases as you go downstream, but the Bradshaw model of River characteristics states that the load particle size should decrease as you go downstream. This will be evaluated late.
My main findings.
My results show:
Fig. 27
The Bradshaw model of River characteristics states that:
Fig. 28
I will now find the similarities and differences between my results and the Bradshaw model of River characteristics:
Fig. 29
Evaluating similarities and differences between my results and the Bradshaw model of River characteristics.
All of my answers are the same as what the Bradshaw model of River characteristics states except for the load particle size and the channel bed roughness.
Why?
The water depth increases due to erosion. The occupied channel width, discharge and wetted perimeter increases due to erosion, tributaries joining, surface runoff into River, and through flow etc. Velocity can be affected due the channel shape, the roughness of the channel bed and banks and the channel slope. The gradient should decrease due to the land flattening as you go downstream. The hydraulic radius will increase going downstream due to due to bigger depth and width of River, so less water is in contact with bed and banks, so there’s less friction.
Load particle size.
The technique used to find the Load particle size was the D90 bedload, this is just one of many of the ways to find the Load particle size. Whilst picking out 50 rocks there could have been favouritism, the smallest and largest rocks wouldn’t have been picked because the small ones would be hard to find and could have fallen down between the bigger rocks, and the larger ones could have been too heavy to lift. Certain rocks might not have been picked because of the wildlife on them. If I did this again I would find a better way of picking out rocks at random, or I could use a piece of apparatus like a bedload trap which sits in the River, and collects the rocks flowing past it.
Channel bed roughness.
I couldn’t get any measurements of channel bed roughness, because I didn’t have any equipment to measure this and there are no calculations to find channel bed roughness from my primary data. The only way I can measure this is by looking at the cross section diagrams, and looking how smooth the channel bed is. From these diagrams site one and two are both quite smooth along the bed of the River, and site three is slightly rougher. But this is not an accurate measure of the channel bed roughness. If I did this again, I would use appropriate equipment to measure this.
Was my hypothesis correct?
Fig. 30
Evaluating the enquiry.
Fig. 31
Overall.
The River Lyn does follow the Bradshaw model of River characteristics very well, apart from the load particle size, which the form of data collection I used was not very accurate. Load quantity and channel bed roughness were not researched, and there was no secondary data suitable for using.
Overall the whole enquiry went very well and I am very proud of what I’ve achieved. I was able to answer all the key questions and nearly answer the question, ‘to what extent does the River Lyn conform to the Bradshaw model of River characteristics?