Geography 4A Skills Paper
Explain the reasons why you selected this aim for investigation Jun 2010 (10)
Our aim was to measure fluvial downstream changes with reference to the Bradshaw model. The aims were met in hypotheses such as that velocity and CSA would increase, whilst channel roughness would decrease with distance downstream. This aim was selected as the variables of velocity, depth; width etc. could be measured using the equipment available to us (flowmeter, meter ruler, tape measure) the depth could also be collected safely, as the river Harbourne’s depth does not exceed 0.5m. This allowed the variables to be collected within 1 day making the investigation time efficient. The methods of data collection were straight forward and allowed more complex analysis e.g. spearman’s rank to be carried out after, in order to see the significance of variables against distance downstream, therefore proving our hypotheses correct. The Bradshaw model also allowed a comparison to a known and consistent trend; that variables increase/ decrease with distance downstream, therefore it was a suitable investigation for varying levels of understanding of the topic and one that a range of students could carry out.
Describe the location of your fieldwork investigation and its relevance to the aim Jan 2010 (12) or Describe the geographical reasons for carrying out your investigation in the chosen area Jun 2012 (10)
I carried out my investigation on the river Harbourne, a tributary of the river Dart in south Devon. The aim of our investigation was to measure downstream fluvial changes with reference to the Bradshaw model. We summarised our aims in hypotheses such as that velocity and cross sectional area would increase downstream and bed load size would decrease. Where variables were not covered by the Bradshaw model out aims were informed by textbooks to find a common theory. The river Harbourne exhibits characteristics of a typical river, as an upland river with its source high on Dartmoor (277m above sea level) with high annual rainfall (1730-2020mm). It exhibits a ‘V’ shaped valley in its upper course and meanders in its lower course, allowing it to be accurately compared to the Bradshaw model. Being a tributary of the river Dart (a considerably larger river) allowed us to carry out the investigation without danger of river levels becoming too high, as the rivers maximum height is 0.5 m at Beenleigh, 18km from the source. This allowed us to effectively meet our aims which were to measure downstream change, by collecting variables from each course of the river, and make a consistent comparison with the Bradshaw model with overall positive results. We were also able to use a method of stratified sampling to select our sites. On a practical level the river was accessible throughout its course, permitting the use of stratified sampling to select sites only a short walk from the road in the upper, middle and lower courses. Therefore we could meet our aims safely and efficiently. We also had secondary data from previous investigations; this was useful in making sure our results were reliable and to extend our investigation.
Where the river experienced channel management strategies, at Harbourtonford 18km from the source, the management was taken into account when aims regarding velocity and discharge were being analysed. In a wider context the drop in bedrock type from granite on Dartmoor to sandstone 3.5km from the source showed the area as a whole was worthy of study.
Select one method of primary data collection that you used in your investigation
Explain the relevance of this method to the aim(s) of your investigation Jan 2011 (8)
Explain one method of data collection that you used was suitable for the investigation Jan 20101 (6)
Outline one method of primary data collection used in the field Jan 2012 (6)
When calculating the discharge we first measured the depth and width of the river to calculate cross sectional area, and then timed this by the velocity. When measuring the width we used a tape measure stretched across the surface of the water, taking a reading from bank to bank. We divided the width by 4 to measure depth at 4 equidistant points across the river, by placing a meter ruler into the river measuring the value at which the ruler met the river bed and took an average. This method was permitted as the rivers depth did not exceed 0.5m. This method was suitable to our investigation because the equipment was transportable when walking from access roads to the sites as well as time efficient, allowing us to sample 6 sites in one day. To measure velocity we used a flow meter, which we placed with the propeller facing upstream at 3 intervals across the width (width/3) for 30 seconds, and timed using a stopwatch and then took an average. This method was suitable as it was simple and quick to carry out, and allowed us to get accurate results for the variables to calculate discharge to compare with the Bradshaw model. We could also compare the individual results by plotting a scatter graph of variables against distance downstream.
Evaluate the effectiveness of this method in the light of your experiences in the field Jan 2011 (8)
Evaluate the suitability of this method of primary data collection Jan 2012 (10)
This method for measuring discharge was effective as it allowed me to collect variables of velocity, depth and width safely and easily in order to compare them to the Bradshaw model. We used equipment available to us such as a meter ruler, which was suitable because the river Harbourne’s depth did not exceed 0.5m and therefore permitted the use of this equipment. Measuring the width using a tape measure was suitable as the rivers width and depth was shallow enough to allow us to measure variables in the river whilst remaining safe, due to the river Harbourne’s being a relatively small river. Measuring the velocity by taking an average of 3 points across the width and between 5 groups gave an accurate average, as we used a velocity meter which timed and calculated the velocity for us. This helped eliminate human error involved in other methods, proving the method effective as it gave us variables to compare to the Bradshaw model and use to calculate other variables e.g. CSA and discharge later.
All of this equipment was transportable from access roads and quick to carry out, making our investigation time efficient as we could collect data from 6 sites along the rivers course in one day.
We could also use a method of stratified sampling to select safe sites from which to collect data from, near to access roads. However I now know this to be biased and that a method of random sampling is more appropriate and effective in getting a truer outcome. On the other hand this was suitable for our investigation due to time constrictions and the safety of our group.
As a result of your experiences in the field justify one or more improvements that you would make to a method of data collection Jan 2010 (12)
Several improvements could have been made to our method of data collection during our fieldwork. Piloting the investigation would have been useful to rule out sites that were not suitable (unsafe or significantly affected by river management), as well as to test the equipment in order to get the most accurate result possible. When collecting variables to calculate discharge more readings for width were needed, as although the river was consistently wider at each site, there were variations within each site, and therefore an average would have been more accurate. When taking an average of a measurement (e.g. width and depth) a more precise method was needed as one wrong measurement can skew the average and any further statistical tests, significantly affecting the outcome of the investigation. Using more precise measuring equipment would also eliminate human error. A limitation of our method was measuring velocity, as the flowmeter had to be precisely calibrated for each site, this could have been improved by using less sensitive equipment.
Describe and justify how you made the data collected were as accurate and reliable as possible (12)
When collecting our data for variables such as velocity and depth in order to compare it to the Bradshaw model, we took averages of 3 equidistant points (velocity) and 4 equidistant points (depth) across the width of the river. This made our data more accurate as it gave a truer representation of the variables. We also took averages of 5 groups of velocity and depth measurements eliminating anomalous results before calculating an average improving the reliability of the investigation. When carrying out statistics tests and plotting graphs we again eliminated any anomalous results (such as a CSA of 0.97 m^2 at Harbourtonford), caused by flood management (Palmer dam).
When calculating velocity we used a flowmeter which measured the time and calculated the velocity for us improving the accuracy of the result over more manual methods. When measuring depth we read the depth reading when the ruler was perpendicular to the flow of water, so that the water did not rise up one side of the ruler causing the reading to be inaccurate.
We used stratified sampling to choose 2 sites in the upper, middle and lower course of the river, in order to get a spread of data from all courses so as to be concordant with the Bradshaw model. However a less bias method such as random sampling would have been more appropriate here. When comparing results to the Bradshaw model where certain variables were not included e.g. hydraulic radius our aims were informed by using a selection of textbooks to find a common theory. We also used secondary data where sites were inaccessible in order to have data from all courses of the river.
When collecting and subsequently analysing variables of discharge and velocity in the lower course of the river, we took into account the river management around Harbourtonford such as Palmer dam, and that settlements may affect discharge due to increased runoff.
Explain why you used one method of data collection and evaluate its effectiveness in showing the data (12)
A method of data presentation I used in my investigation was a scatter graph, as it allowed me to plot continuous numerical data to show whether a correlation existed between the two variables (or a variable against distance downstream) and how strong this correlation was. It was useful as it was easy to see if a correlation existed e.g. when plotting CSA against velocity I could that both increased downstream [0.013m^2 , 0.081 m/s on Harbournehead] , [0.74m^2, 0.2m/s at Harbourneford] and [0.71m^2, 0.82m/s at Beenleigh]. It also allowed me to easily see anomalous results and emit them from statistical analyses. I could plot a line of best fit enabling me to predict variables in further sites, as well as test how significant the correlation was between two variables using statistical techniques such as spearman’s rank.
A limitation of this technique was that I could not plot discontinuous variables such as bed load roundness, therefore I couldn’t analyse all the variables against the Bradshaw model and analyse all the variables in to the fullest. A bar graph would have been more appropriate here, however, this made the comparison to continuous data more difficult. However a scatter graph was the most suitable for my investigation as it had benefits of other techniques. Such as I could use it with the majority of my data and showed simply whether a correlation did/didn’t exist. I could read off data easily with two variables shown on one graph, not possible with e.g. a bar graph.
Measures taken to minimise health and safety Jun 2012 (8)
In order to combat the risks involved I researched the river and surrounding area before undertaking the practical work. We collected our data in February when there were high levels of precipitation and particularly cold weather. Therefore I felt hypothermia was a significant risk, especially at site one on Dartmoor as there is little shelter and high wind speeds. To overcome this I made sure my group wore suitable waterproof clothing (e.g. waterproof gloves for insulation) and timed using a stopwatch the amount of time spent in the water, we also checked weather reports to know when to expect rain. I also made sure my group wore warm clothing and took spare clothing with them.
I also felt slipping and falling was another risk due to the potentially wet weather and harsh environment particularly in the upper course. To minimise the risk I made sure my group wore walking boots, and took care collecting data in and around the river.
Another risk was that from oncoming traffic, as many sites in the lower course were near busy roads. Therefore high visibility vests were worn and care taken when crossing busy country roads.
Due to the high levels of precipitation before and during our trip, and in the upper course in particular I continually monitored the potential risks including rising river levels above 0.5m.
Explain a technique of data analysis and what made this appropriate (12)
I used spearman’s rank correlation coefficient to analyse my data and to objectively prove there was a correlation between decreasing gradient and velocity, which there was. My result gave a coefficient of 0.94 which was significant at the 5% (0.88) and 2% (0.94) significance level, therefore the null hypothesis could be rejected and I could accept that as gradient decreased velocity increased concordant with the Bradshaw model. This technique was appropriate as I could examine the strength to which two different variables were related. It allowed me to test the significance in order to prove that the relationship was not by chance (by looking at its significance) so that it could be compared with the Bradshaw model. It was more suitable over other methods such as Chi squared because I could test the relationship of two different quantitative variables which most of my data was, however as the chi squared test deals with grouped qualitative data it would have been more useful for analysing bed load roundness.
Spearman’s rank enabled me to find how much that data was correlated between -1 and 1 giving an immediate indication of the strength of the relationship. The test was suitable for the number of values I had collected for each variable as opposed to e.g. chi squared where the total number observed had to exceed 20.
The technique also allowed me to compare variables were one was dependant on the other e.g. velocity is dependant on gradient, as many river variables are dependant on others (CSA, bed load roundness, bed load size and velocity) therefore it was a suitable as the test allows for this. Comparatively Mann Whitney U requires the variables to be independent of each other therefore spearman’s rank was the appropriate choice.
With reference to your aim evaluate the success of your investigation Jun 2010 (12)
The conclusions that were made from my investigation were very useful in understanding the underpinning theory of the investigation that river variables changed downstream. The Bradshaw model has also become much clearer to me as a result of all my hypotheses being correct, and therefore proving my investigation a success.
By using equipment such as a flow meter, dumpy level and tape measure I became familiar with it by measuring river variables. I was also able to see a positive correlation between cross sectional area and velocity as you moved downstream e.g. [0.013m^2 , 0.081 m/s on Dartmoor] , [0.74m^2, 0.2m/s at Harbourneford] and [0.71m^2, 0.82m/s at Beenleigh], again proving the success of the investigation. Added to this I could also see that channel roughness decreased from source to mouth in line with the predictions of the Bradshaw model.
My understanding of river variables was aided with the use of graphical techniques of data presentation, and furthered my understanding of how the variables change and the features located along the course of a typical river. This varied from small tributaries and saturated ground at the source - due to the impermeable granite bed rock, to an overall larger more powerful river near the confluence changing to sandstone bedrock.
The use of statistical techniques including spearman’s rank and Mann Whitney U also proved the usefulness of the investigation due to the spearman’s rank result of 0.94 which was significant at the 5% and 2% significance level. This meant we could reject the null hypothesis and accept there was a correlation between gradient and velocity. This also extended my knowledge on analysing data, and the importance of an objective test when drawing conclusions. In wider context I have also learnt about the safety precautions necessary when undertaking working the field. Therefore my investigation again proved successful.
On the other hand I have also understood the improvements that need to be made, for example that more readings need to be taken and averages calculated to get a more accurate result. But to also consider the human impacts such as selecting sites near access roads, where there may be increased runoff due to impermeable surfaces. Therefore I should have used a method of random sampling to eliminate bias. However carrying out the investigation first hand, also allowed me to see the implications of river management. For example the Palmer dam at Harbourtonford protects the town of Harbourtonford from flooding but also affects discharge downstream, which could impact ecosystems in both the Harbourne and the river Dart due to reduced silt sediment in the lower course.
Assess how useful your investigation was to your understanding of geography/geographical theory Jan 2011 (12)
My investigation was to measure the downstream fluvial changes of the river Harbourne and compare them to the Bradshaw model. My investigation greatly improved my understanding of the fluvial changes of a river as all my hypotheses were proved correct; therefore my investigation was a success. I could see first hand the fluvial changes downstream, in that as velocity increased CSA also increased ([0.013m^2 , 0.081 m/s on Dartmoor] , [0.74m^2, 0.2m/s at Harbourneford] and [0.71m^2, 0.82m/s at Beenleigh]) as well as other variables such as hydraulic radius, width and depth. I could also see that gradient and bed load size decreased, concordant with the Bradshaw model and was able to better understand these changes.
I have understood the need for sampling when selecting sites, as well as the use of stratified sampling in order to select safe and accessible sites. However I understand now the need for more random selection as by using stratified sampling I left my investigation open to bias.
The application of statistical tests such as spearman’s rank not only improved my understanding of the aim by the hypotheses that gradient decreased with velocity increase being proved correct (significant at the 5% 0.88) and 2% (0.94) significance level result (0.94)), showing that there was a correlation between CSA velocity. But I also became familiar with using statistical test and understand their importance in giving objective proof/disproof. I was able to understand how downstream changes were linked as the data I collected supported the predictions it made. I can now appreciate the health and safety risks posed my carrying out practical investigations
By carrying out practical work I have been able to see the uses of the Bradshaw model and where it did not fit with characteristics of a real life river. Evident when collecting data at Harbourtonford where variables such as velocity were uncharacteristically low (0.7), due to the flood management of palmer dam. I could consider wider implications of this, such as that disruption in velocity and sediment flow can have negative ecological effects further downstream (and possibly into the river Dart). The investigation helped me to understand the general nature of the Bradshaw model as it a suitable comparison of most characteristic rivers. However even though my results show a concordance with the model there is only a slight agreement with many variables. Therefore I have learnt that whilst theories are useful in making loose predictions they are not a basis for judgement in real life.
Evaluate the usefulness of you conclusions and consider potential implications Jan 2010 (10)
In the light of my results I can see a correlation between downstream change and the increasing/ decreasing of specific river variables such as with velocity and CSA [0.013m^2 , 0.081 m/s on Dartmoor] , [0.74m^2, 0.2m/s at Zempsters bridge] and [0.71m^2, 0.82m/s at Beenleigh]. However I could also see the affects of river management as the mean velocity dropped uncharacteristically to 0.66 at Harbourtonford (site 5). This could be a further extension of our investigation as such variables of velocity and sediment size and load can affect the ecology of the river further downstream. A point of development that could be made is to compare our results with secondary data from previous investigations along the Harbourne during the same period (February) to see how the river has changed over time and why, for example extensive farming around the area or building of new roads/ town which could affect discharge due to increased runoff caused my impermeable surfaces. As well as comparing our results with those collected at different times of the year to investigate whether the river Harbourne fits the model at different times of the year, and why (e.g. increased vegetation may catch sediments in the summer months decreasing load quantity).
