Lucy Barker – Historic Hydrological Droughts


Hello my name is Lucy Barker, I’m a
hydrologist at the Center for ecology and hydrology in Wallingford and I’ll be
talking today about historic hydrological drought data and the
hydrological drought Explorer so firstly we’ll maybe think about what is the
drought and just generally occur when there is less moisture available than
expected and this to be deficits in moisture can occur when we’ve got
deficits in rainfall in soil moisture and groundwater and streamflow all
leading to different types of drought and today what we’re really interested
in is hydrological droughts and that’s when we have lower river flow than we
would expect for that time of year why are we interested in historic droughts
and especially when no two drought events the same in terms of their timing their
characteristics and the spatial extent that they affect well and we need to
understand the past variability to understand and detect changes in the
future and we also need to understand historic droughts and they because that
helps us to inform planning and management for future events and for
example this is particularly important in UK drought drought planning from the
water supply sector where historic droughts and the worst droughts on
record are used to inform their drought plans there are lots of different ways
to characterize droughts and historic droughts and the way that we chose to do
this was to use the standard stream flow index this is a drought indicator
which is one of many and it’s part of a family of standardized drought
indicators all based upon methods derived for precipitation and these
methods all use and the same kind of calculations which basically give you
the standard deviation from the long term mean and basically showing whether
it’s wetter or drier than normal for that particular month
so we have negative values when it’s dryer or there are flows are lower than
normal or we have positive values when it’s wetter than average or flares are
higher than normal and there’s an indicator values that can be associated
with severity categories but the more more negative the values the more severe
and they’re less likely that is to occur some other benefits of using the
standardized streamflow index and these other standardized route indicators is
that they’re comparable over time and space because we have different
variations for different variables like rainfall and groundwater we can compare
the drought in characteristic across these different variables and these
different parts of hydrological cycle for example you could look at the standardised precipitation index on the UK drought portal which is available
at this link which is updated in near-real-time and allows you to look at
rainfall deficits across the UK and we’re looking to add streamflow and
groundwater dates this so you can really monitor and droughts across the UK in
the holistic way and that deficits in rainfall river flow and groundwater in a
comparable way and one really great thing about these standardized
indicators is that and they you can calculate them for different accumulation
periods and whether that’s the one month so you’re looking at kind of
instantaneous deficits in flow or precipitation or whether you’re
interested in longer periods maybe like three months if you’re interested in
seasonal deficits up to six 12 18 24 months and deficit depending on the type
of drought you’re managing and then all the resources that you’re managing as well
and as I mentioned they indicate the severity and the probability of low flow
and these indicators their SSI initiated rainfall indicators really used quite
widely internationally but hasn’t been used too much here in the UK before and
so we’ve been doing some work over the last few years which and look at the
best ways of calculating these in the UK and what they
shown in terms of about characteristics and drought propagation using observed
data but these observed data which we hold on the National River Flow Archive
generally only start in 1961 and we know in that period before 1961 there was
some pretty severe that events which effects had a wide-ranging impact on the
environment society in water supply so within Historic Droughts we’ve been
looking at… we have derived the standardized streamflow index from 1891
to 2015 calculate these data we used historic reconstructions of daily
rhythms like for 303 catchments which Katie Smith is produced and those data
are available to download for free from the environmental data and Center and
there’s also a video you can watch where Katie and describes the modeling process
fanchetti how to download the data and explore this so but the standardized
stream flow index all these 303 catchments is also a thing available to
download from the e IBC and I’ll show you how you can download that later on
but so this this time series show is the deficits in flows or the flows from 1891
to 2015 but if we want to look at the individual drought events we first have to
extract those events from this time series and the way that we did this was
to define a draught event as a consecutively negative period of SSI
values with at least one month reaching or exceeding a given threshold we’ve
used three thresholds to extract events – one which equates to a moderate
drought – 1.5 equating to a severe drought for –
to equate into an extreme drought and if we use the minus 1 threshold we’re gonna
maybe identify quite a lot of droughts because it’s… they’re going to be
occur more commonly than events where with if the value reaches
– 2 because that’s much less likely to occur for each of our extracted events we can
calculate a range of event characteristics we have the duration simply just
a number of months within that event the accumulated deficit which the sum of the
ssi values within that event we can look at they cumene deficit the humors with
deficit divided by duration which gives us an indication of whether it was a
short sharp event or kind of a very prolonged event that is maybe just
rambling on and not too severe that’s lace and we have the maximum intensity
which just gives us the minimum SSI value within that event these extracts
and events can all and their SSI time series for the 303 catchments can all be obtained from the UK Hydrological Drought Explorer which I will show you
in just a second and is available at the link on the
screen and if you want to find out more about the model flow constructions you
can look here at the reconstruction of the data rates or a which tells you
gives you information about how well the models are performing in each catchment
which might be which is particularly important when thinking about using
model data and we want to know that the model data is actually representing the
observation as well so before before you dive in and use any of the data I really
recommend that you have a look at though the reconstructed flow for your selected catchment so next I’m going to show you how you can download the
data from the EIDC – the Environmental Information Data Center and you’ll need
to log in and register for the service which is entirely free once
you’re logged in and if you go to the data page here and click download data
you’ll get a page here which lists and the files available which is one for
each of the 303 attachments and they are identified by the National River Flow
Archives ID if we go down we can select attachment
unload it and we can see here in the CSV we have
columns for the seven different accumulation periods available 1 3 6 9
12 18 and 24 months and because these accumulation pairs are ugly expects a 10
month accumulation period and we have some na or – 9 9 9 values so for the 3
months accumulation period the first-two
values of the time series are n/a they don’t exist because they’ve been used to
derive the SSI value for March 1891 and similarly if we look at 24 month
accumulation period we have 23 mas here before we get to our first SSI value in
December 1892 also on the data page and we have some information on how the data
were derived and you can also download some supporting documentation which will
give you a lot more information on on the calculations used to derive the beta now to the exciting bit where we can
look at the UK Hydrological Drought Explorer app which allows you to look at
the SSI time series and the extracted events for all of the catchments so on
the first page here we have a map of the 303 catchments from which data are
available on the left here you can choose which catchment you want to look
at and the map will update to show you where that catchment is once you select
your catchment you can go to the catchment plots page where you will be
able to look at the time series plot of your selected catchment and accumulation
period you can change the accumulation period and the map will update and you
could change the time period you want to look at it’s the default is for it to
look at the four times there is 1891 to 2015 but you could look at and if you 10 1950 and the graphs
will update to just show you that period if you scroll down you can then graph
the load shows simply just where the extracted drought events occur in that
time series and if we remember and the drought event was defined where have-
SSI values with it with at least one month reaching our given threshold so by
default we’re showing the severe droughts right with a threshold of minus
1.5 and we can see here we have quite a lot of droughts identified within
this 1910 to 1950 area for us an attachment then at the bottom of this
page we can look at the extraction offense but with peer plotting and with
the time on the x-axis that we’re sharing duration on the y-axis so we’ve
got a very long event here in the early 1940s and the size of dots is the
maximum intensity lowest SSI value within the event and
they are coloured by the mean deficit so it was quite a long event it was maybe
in terms of the SSI values per month is maybe not so severe compared to this
event right down in 1929 which was quite short but obviously have some very
severe SSI values in there so we can look at all of this information we could
look at different patch meant and look at the same information for selected
time period accumulation peri events threshold next leap and look at how
those extracted events ranked in terms of the different different
characteristics so for your selected catchments accumulation period and event
extraction threshold we are ranking the events by duration and showing the ten
most severe events so if there are selected catchments in the northeast of
England the most severe event in the longest event on record between 1891 and
2015 was in 1995 to 1998 whilst if we ranked by accumulated deficit things
maybe haven’t changed so much but we can look at another characteristic being
deficit and that ranked and painless quite substantially and in fact most of
their event now is 1955 to 1958 and but at the bottom underneath the table is
just showing you where well it’s about current being used to feel
selected characteristic and then carry on being the ranked tenth so we’ve been
looking at individual catchments so far but we now are going to look at things
on a more regional scale on the regional analysis tab and the graph might take
just a couple of moments to load for you and for your select accumulation period what we’re looking at here is for each
region was shown on the map on that introduction page shows the centre of the model area that’s selected any one time and we have nine regions across the UK so when
a hundred percent of the modeled area of each region is in drought at the same
time our bars would go up to 900 on the y axis and for example for Northern
Ireland here the green at the bottom you can see if actually a hundred percent of
the modeled area is in drought and several times throughout our full period
of record here and this graph is exciting allows you to zoom into a
period you’re interested in we’ve got here in the 1930s and 1940s we have
almost a hundred percent of each of the modeled areas and in drought in 1929 in
the mid 30s and also in the late thirties though you can hover over and
it’s going to tell you the percentage modeled area and drought in any one
month and we change the accumulation period we make more interesting deficits
just in the one month scale and you can see here in ninety a start of 1929 we
have almost a hundred percent of the country in drought at any one time
time consumed in and see how that changes maybe by the regions and before
and after that and very intense event there we also has some summary maps at
the bottom which has plot up the regions for the start month of seasons
for each year in the series and you can scroll through and see how maybe that
changes and across the year and more more in a map the information we’ve
shown so far has been taken from the best model run of the reconstructed flow
ensembles but because we’re looking at model beta this a lot of uncertainties
associated with that and what we wanted to do was look at the uncertainty
involved associated with the model reconstructions and how that affected
the SSI confirmations so for a handful of catchments what we’ve done is we’ve
calculated the standard stream flow index for the top 500 runs from the
ensemble for a subset of accumulation periods we’ve only used a few catchments
so far and a subset of the accumulation periods there’s quite a long time to run
and we’re currently setting this up for for the 303 catchments and some more
accumulation periods which will be running on supercomputers so it will run
a lot faster than money on our desktop machines and but for these study
catchments and for your select accumulation period this is what this
graph is showing you is the reconstructed SSIS in the turquoise
compared to the SSI as calculated from the observed data observed later from
the National River flow archive and again this graph allows you to zoom in
to a time period so if we look here where the observations start for
the Avon and we can see that the best model run in the dark turquoise here and
the pale shading behind it is the range of the ensemble and we can see how that
compares to the observations in the orange as they through time you can look
at a different accumulation Paris because the uncertainty might be
different and missed it and how their model performs might change the representation of the as the model
data competitive observations and maybe see how different catchments perform and
whether there there is more or less uncertainty and when seeing this
groundwater dominated catchments in the Southeast of England the the range of
the ensemble is much narrower so maybe there’s less uncertainty with the
modeling but maybe how it compares to the observations as well so there’s a
lot of things to take into consideration but if you’re interested in how the
models perform and I really recommend that you look at the reconstructed Flow Data Explorer which you can find a link to on the About Drought webpage at the end
here and as well as link to download the reconstructed flow data here or the
standard stream flow index see you throughout this part so I hope you’ve
enjoyed this quick run through of and the hydrological drought matrix flora
and the standby stream flow index pages that are available to download if you
have any other questions on anything you’ve seen or I said today we let me
know and I’ll be happy to get back to you thank you

Leave a Reply

Your email address will not be published. Required fields are marked *