1 of 5

GROUP 12

2 of 5

  • How many years of data do we have?

Flow (Q, q): 18 Years (and 2 months+, 10/10/1991 to 12/29/2010) (19 years, 2 months, 25 days)*

Rainfall: 39 years + (1/2/1984 to 11/28/2023), Jesmond

  • Is the data complete? Are there any missing values?

Few missing values in Q data and Rainfall (long_Meadows)

99.15% data availability

  • What are the units of the data?

m3/s (Q), l/s*km2 (q), mm (Rainfall)

  • What is the temporal resolution of the data?

15 minutes

Task 1: Basic stats

​

3 of 5

Task 2: Hydrology

​

Mean: 0.3160 (Q), 5.747 (q, unit flow), 0.020 (Rainfall, Jesmond)

Median: 0.122 (Q), 2.22 (q), 0 (Rainfall, Jesmond)

Std Deviation: 0.758 (Q), 13.779 (q), 0.131 (Rainfall, Jesmond)

​

Plot two years of data – 94/95 and 07/08

​

​

    • Are there seasonal variations? Wet and dry seasons in Newcastle?
    • Are there a lot of inter-annual variations? What are the differences between the 94/95 year and 07/08?

Plot example: Wet (15/16) and dry (95/96) years flow values for a station in Cumbria

4 of 5

Task 3: Extreme Hydrology

​

Quantiles

​

​

​

Even more extreme!

    • Annual maxima – compute the maximum flow for every water year in the timeseries – a water year in the UK starts in the 1st of October at 09:00 AM

​

​

    • Highest ever recorded flow values = 25.3 m3/s
    • Highest ever recorded rain values
    • Median of the AMAX series of flow = 10.35 m3/s
    • Median of AMAX series of rainfall =

AMAX series

5 of 5

Data Quality Control

Imputation – how to treat missing data:

​

CHOICE : Keeping all values, because we need a bigger record for finding the peak rainfall with a bigger return period. Imputation by mean and further finding the spatially distributed rainfall by using Thiessen method.

Reason : The rainfall for the overlapping period (period where data is available in all the stations) was checked. It is found that the use of the thiessen method for interpolation with imputation by 0s resulted in values that were less than the values obtained by imputation by mean.

Calculation of mean rainfall for spatial distribution and use of mean values for imputation gives slightly lower values. In a similar way the use of median values for interpolation resulted in under-estimation of the values.

Moreover, adding 0s in place of unknown values is not practical.

→ Thiessen method gives better spatially distributed rainfall values.

How much of the different timeseries to keep?

​

DATA : from 1/2/1984 till 20/6/2018