Advanced

Den förorenade jordens gåta

Thedin, Rasmus LU (2017) STAH11 20162
Department of Statistics
Abstract
When analyzing contaminated soil today, the method used consists of taking five samples from an area of 20 times 20 metres. Then UCLM95 (95 per cent upper confidence level of the mean) is calculated and compared to critical levels. The issue with this specific method becomes evident when the results often come out inaccurate, which leads to an assumption that the soil is not as contaminated as it actually is. Either it gets underestimated, leading to decisions that might be bad for the environment, or overestimated, leading to decisions that cause larger expenses than actually needed.
The aim for this paper is to find out how good the precision of the method used is. This is made with simulations in R. The paper also tries to find out how... (More)
When analyzing contaminated soil today, the method used consists of taking five samples from an area of 20 times 20 metres. Then UCLM95 (95 per cent upper confidence level of the mean) is calculated and compared to critical levels. The issue with this specific method becomes evident when the results often come out inaccurate, which leads to an assumption that the soil is not as contaminated as it actually is. Either it gets underestimated, leading to decisions that might be bad for the environment, or overestimated, leading to decisions that cause larger expenses than actually needed.
The aim for this paper is to find out how good the precision of the method used is. This is made with simulations in R. The paper also tries to find out how many samples that are needed to get a satisfying precision.
A small section of the paper also looks at what would happen if, instead of using UCLM95, the mean was to be used.
The results indicate that with method used today, the soil is catagorized wrong more often than right, but due to the construction of UCLM95, overestimation is more common than underestimation, which leads to economic consequences rather than environmental ones. The results also show that, if you want to have a satisfactory precision (in this case 80 per cent), the sample size would have to be increased, in relation to the amount being used today. This would, however, not be economically realistic due to the requested size of the sample needed.
It is also shown that when using the mean as comparison to the critical levels, 30-40 samples would be enough, and could be used with cost effective methods. (Less)
Please use this url to cite or link to this publication:
author
Thedin, Rasmus LU
supervisor
organization
alternative title
Hur väl fungerar dagens metod för klassning av förorenad jord
course
STAH11 20162
year
type
M2 - Bachelor Degree
subject
language
Swedish
id
8900248
date added to LUP
2017-04-12 10:14:44
date last changed
2017-04-12 10:14:44
@misc{8900248,
  abstract     = {When analyzing contaminated soil today, the method used consists of taking five samples from an area of 20 times 20 metres. Then UCLM95 (95 per cent upper confidence level of the mean) is calculated and compared to critical levels. The issue with this specific method becomes evident when the results often come out inaccurate, which leads to an assumption that the soil is not as contaminated as it actually is. Either it gets underestimated, leading to decisions that might be bad for the environment, or overestimated, leading to decisions that cause larger expenses than actually needed.
The aim for this paper is to find out how good the precision of the method used is. This is made with simulations in R. The paper also tries to find out how many samples that are needed to get a satisfying precision.
A small section of the paper also looks at what would happen if, instead of using UCLM95, the mean was to be used.
The results indicate that with method used today, the soil is catagorized wrong more often than right, but due to the construction of UCLM95, overestimation is more common than underestimation, which leads to economic consequences rather than environmental ones. The results also show that, if you want to have a satisfactory precision (in this case 80 per cent), the sample size would have to be increased, in relation to the amount being used today. This would, however, not be economically realistic due to the requested size of the sample needed.
It is also shown that when using the mean as comparison to the critical levels, 30-40 samples would be enough, and could be used with cost effective methods.},
  author       = {Thedin, Rasmus},
  language     = {swe},
  note         = {Student Paper},
  title        = {Den förorenade jordens gåta},
  year         = {2017},
}