The River Brent

Our Investigation

To what extent is Bradshaw’s model accurate over a small stretch of the river
- We have learned what Bradshaw’s model says about how rivers change from source to mouth
  - We know that his model is mostly correct

- We wanted to find out whether it is also accurate over a short stretch of river and if not, what were the factors that may have affected the river
We started our investigation in the Dollis Brook in Hendon (a tributary of the River Brent) and followed the stream until after its confluence with the River Brent, so part of our investigation was in the River Brent
- The River Brent is itself a tributary of the River Thames
- We did our investigation here because it is near our school so we can access the site cheaply

Risk Analysis

It is important to be safe in the river
- The river is generally quite shallow but can get deep at parts and be strong after heavy rain
Ensure that we have safe access to the river
Wear suitable clothing including well sealed and gripping shoes
Before entering, put a ranging pole in first to check the water isn’t too deep
Always put the ranging pole in front of us before walking forward wherever the river is too deep to see the bottom
Don’t go into the river where the water is deeper than 1 metre
Don’t go in the water if the water is flowing very fast
Don’t go in the water if we have an open wound because it can get infected
If the water appears too polluted don’t enter
Help each other in the water to ensure your friends don’t slip
Have a phone (kept out of the water) so that you can call for help in an emergency

Our Sites

We used systematic sampling to choose the sites
- Each was 500m apart and we investigated 5 sites so in total we went over a distance of 2km
Site 1 (start point- shallow stream)
Site 2 (500m- small stream but some human intervention with concreted banks and a small weir)
Site 3 (1km- there was a meander here)
Site 4 (1.5km- this part was much wider and completely altered by humans with concrete banks and bed)
Site 5 (2 km- much narrower site but more natural and just past the confluence with another tributary)
We worked in small groups to help each other out, and to be able to take measurements when the method required more than one person.

Our Methodology

To measure	Method and Justification	Evaluation
Width	Using a tape measure, we stretched it across the river from one bank to the next. We recorded the reading in centimetres. We needed to see whether the river gets wider downstream.	It was important to make sure that the measurement went at right angles across the river bank. As different parts of the same area of river had different widths, we had to choose where to take the measurement which meant it wasn’t reliable. We had to hold the tape measure tightly to take the reading more accurately.
Depth	Used a metre stick, we used systematic sampling to take measurements every 50cm across the river. We did this so that we would be able to find the average depth of the river in different parts of the river.	We had to make sure that the metre rule was held upright (perpendicular to the river bed), and not at an angle otherwise our reading was inaccurate. We took measurements every 50cm because this was manageable and is probably fairly accurate. However, since we didn’t measure all the samples, we don’t know the depth between our samples which could be very different, so our average depth may not be accurate. Also, other parts of the same site may have different depths.
Velocity	We measured out 10 metres with the tape measure along the length of the river in the middle of the channel We held a ranging pole at each end to mark out the 10 metres. We dropped a dog biscuit upstream and timed with a stopwatch how long it took to get to the other ranging pole. This helped measure how quickly the water was flowing. We used dog biscuits because they float, are cheap and biodegradable. We repeated the method three times at each site to get a more accurate answer.	In some places the biscuit didn’t move or got stuck, so we discounted these as anomalous results We had to be careful not to move while we did the experiment, so we didn’t create currents in the water. The biscuit did not follow a straight path and sometimes drifted into a faster or slower stream of water, so our results weren’t very reliable. 3 repeats were probably not enough for an accurate average. A flow meter would give more reliable and accurate results and allow us to take measurements for different parts of the river or for below the surface, but flow meters are very expensive.
Gradient	We held two ranging poles at 10 metres apart along the length of the river. We used a clinometer to find the gradient by holding the clinometer at a fixed point on the ranging pole and lining up the marker on the clinometer to the matching point on the other ranging pole. We read the gradient from the reader.	It was quite hard to make sure to match the same points on both ranging poles, so our results were not very reliable. Many of us tried but our results were not very consistent.
Stone Sizes	We scooped up a tub full of sediment from the river bed at a quarter, half and three-quarter way across the river. We tried to use random sampling to pick five stones without looking, and used a ruler to measure the longest edge of each stone in mm. This helped measure whether the stones change size as we move downstream.	It was not actually a random sample because very large boulders were not scooped up in the bucket and tiny silt was not picked up. To resolve this, we also did the proportion of different sediment size study mentioned below. Measuring the size of the stone by measuring the longest size was not completely accurate as some stones are long and thin and others very round. Weighing the stones would have been a more accurate way of measuring the size. At some points there were no stones and at some points the water was too deep to reach down for the sample.
Proportion of Different Sediment Sizes	We used a plastic tub and scooped up some sediment from the bottom each quarter way across the river width. We then judged what percent was: silt, sand, rocks that are 1-5mm, rocks that are 5.1-10mm, and 10.1mm+. We used a tub because it was easy to tell the proportions.	It was just estimation about the percentages, so it was not fully accurate. However, we made sure at least three or four people were doing it so that we could use the average of our opinions so that it was more accurate.
Stone Angularity	We compared the stones (chosen above) with a “roundness index” chart. The roundness index showed pictures of six stones with varying degrees of angularity, one was the most angular and six was the smoothest and roundest. Underneath the pictures it described what the stone would look and feel like. We used this because it is visual so you can have a look and get a good impression. This helps us to see whether rocks get more rounded as we travel downstream.	Sometimes we had different opinions so we had a vote to make sure we recorded what most of us thought, which would most likely be the most accurate. However, there are still the problems caused by the fact we may not have sampled randomly. At some points there were no stones and at some points the water was too deep to reach down for the sample.
Field Notes and Photos	We made field notes about the surroundings of the river and the river banks. For example, whether they were natural banks or were built up with cement/concrete. We photographed any interesting natural features like the meander, river cliff and slip-off slope. We did this so we would remember more about the sites to help us to explain our findings.	Sketches were not always so accurate, so we took photographs instead. To remember which site our photos were from we took a photo of Site 1 etc. before taking photos of that site. Field notes helped us with our evaluation. Because many of us were taking photos and field notes, we had lots of evidence.

Data Presentation

We made cross sections of the river at each site
- We were careful to use the same scale for each cross section so we could compare them effectively
- This helped us to compare the cross-sectional area at each site and to see whether the shape was symmetrical or deeper at one side than another
- It was very helpful as it was clearly visual and easy to see differences in depth, width and volume
We made bar charts to compare many of the factors from different sites, for example average stone size, average angularity, etc
- This was good because it helped us to compare clear differences between the means
- One weakness was that it ignored the range of data we collected
We made proportional stacked column bar charts for each site to show the proportion of different sediment sizes we took at each reading
- We put these over the cross sections we had made
- This was very effective because it was very clear and easy to compare and putting them on top of the cross sections made it easy to see whether there were patterns between the different sites or the middle and the edge of each site
We used box and whisper graphs for the sediment sizes we collected from the middle of the river at each site
- This allowed us to compare the median values and find the interquartile range which is a more useful way of examining trends as it ignores the outliers
We used a scatter graph to compare the size and angularity of each piece of sediment we measured
- We put the data from every site on one graph so that we had a big sample size
- We did this graph so that we could see whether there was a correlation between these variables
- We were expecting that because of abrasion and attrition, the smaller the sediment the smoother it would be
We annotated out photographs with things we had noted down in our field notes
- We arranged them so that all photos from one site were together
- They were a useful visual record of the features of each site and any human influences (such as concreted banks) which was very useful when analysing our data

Our Results

Width and depth did not follow Bradshaw’s model
- Although the width was smallest at the start, the width and depth varied a lot from site to site
- This could be partly due of a lot of human intervention along the way
- There were concrete banks at site 2 and concrete banks and beds at site 4
- This doesn’t explain all the differences.
Velocity
- The water was quicker at the last 2 sites
- Site 1, 2 and 3 were all very similar but showed slight increases between sites
- Once anomalous results were discounted the three results taken at each site were fairly like each other
Discharge
- The discharge did increase as the river went downstream, particularly at site 4, where the Mutton Brook has joined the river
- However, there were many parts of the river between our sites which appeared to have low discharge
- For example, the river in Brent Park became narrow and shallower between sites 4 and 5, (There was a lake alongside the river which may have diverted some of the water)
- By the time we reached site 5, we had passed the lake and the discharge was very high
Gradient
- There was very little difference between the sites
- There were very low gradients at each site
Stone size
- The stones at sites 1, 2 and 3 were significantly larger than sites 4 and 5
- However, there were many anomalies and we also ignored places where we couldn’t get samples
- Sites 4 and 5 had far more human management which may have influenced the sample sizes, so it was not a completely fair test
Proportion of different sizes
- There was no pattern at all in the proportion of sediment size either between sites or across sites
- At every site there was a range of samples from only silt and boulders to a mix of all rock sizes
- This could be a clearer indication of the rock sizes than the results from the stone size test
Stone angularity
- There was little change in angularity downstream with very similar values averaging 3-5 at all sites
- There were very few very angular or smooth stones and average values did not follow a pattern

Conclusions

Width and depth (cross sectional area) did reflect Bradshaw’s model in the small area we checked
- There were some areas between our sites that appeared to vary so smaller intervals could give more accurate results
Discharge did match Bradshaw’s model over the short area we measured
- It could have been different in Brent Park and this should be investigated
- Our result though does make sense as a tributary joins the river as well as run off and there were small pipes with drainage water flowing into the river at points along the river too
Gradient results did not show any trend
- This was probably too small an area to notice differences in gradient, but our results may also have been unreliable and inaccurate
Stone size results were inconclusive
- The 5 stones we took at points across the river at each site did suggest that the stone sizes got smaller but examining the sediment we scooped up to get these stones did not show any pattern suggesting that the stones we picked apparently randomly may not have been representative.
Stone angularity does not appear to have matched Bradshaw’s model over this small distance, there was no pattern in our data
Overall, Bradshaw’s model does seem to have been reasonably accurate over a short distance in describing changes to the cross-sectional area of the river, velocity and discharge
Changes to load size and shape were not evident however over the short section of the river we studied.

Evaluation of our Investigation

Most of the results we took were reasonably accurate as we were working in teams and double checking our results
The investigation was well planned and the tests we did went according to plan
We followed our health and safety rules, and everyone remained safe
Considering our time restrictions and limited budget, the investigation was successful and provided a good sample size for studying and quantitative data to produce graphs and analyse
However, there were ways we could have made our results more accurate
- Use an electronic method to measure gradient
- Use a flow meter to measure velocity
- Use scales to weigh stone size rather than a ruler
- Used a set of sieves to measure the proportion of different stone sizes instead of estimating
There were also problems related to reliability
- We only went on one day to the river and this may not be typical
  - We should have many visits to be sure of our results at different times of the year and different weather conditions
- Perhaps 5 sites were not enough
  - We did pass places that did not appear to match Bradshaw’s patterns. Choosing smaller intervals would make for a more thorough investigation
- Each site had different cross sections we could have chosen
- Taking 3-4 cross sections and sets of stone samples at each site could have made our results for each site more reliable