SACNASP CPD EVENT
Well done and Thank you to the GWD Gauteng Branch Chair Mr Kwazi Majola for making this sharing & learning opportunity possible!
BACKGROUND TO THE TALK: Alongside the effects of climate change and anthropogenic factors, natural climate cycles have considerable impacts on the hydrologic cycle. In this study, we look at how global climatic oscillations cycles, like El Niño–Southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD) affect total water storage and groundwater storage in the Orange-Senqu River Basin by analysing two large aquifers: the Stampriet Transboundary Aquifer System (STAS) shared between Botswana, Namibia and South Africa, and the Karoo Sedimentary Aquifer shared between Lesotho and South Africa. The findings could help decision-makers prepare more effective climate-change adaptation plans at both national and transboundary levels.
ABOUT THE PRESENTER: Tales Carvalho-Resende has more than 7 years of experience in the development and management of international cooperation projects on environmental issues, climate change and transboundary waters. He worked at the UNESCO Intergovernmental Hydrological Programme (IHP) where he coordinated and supported research and capacity-building activities on groundwater and climate change, water diplomacy, conflict resolution, and international water law that lead to the establishment of the first arrangements for the governance of a transboundary aquifer in Southern Africa (Stampriet Aquifer in 2017 – Botswana, Namibia and South Africa) and Central America (Ocotepeque-Citala Aquifer in 2019 – El Salvador and Honduras). He is Brazilian, PhD, Earth Scientist and holds an MA in International Affairs and LLM in Climate Change Law and Policy.
We appreciate his excellent presentation and the encouragement by Dr Tales for sharing of the materials.
In vain have you
if you have not
imparted it to others.
(Live questions and responses to these questions are transcribed and might contain capture errors. We will continue with the quality check and update the post accordingly.)
Q: Thanks Tales for the wonderful presentation. It is always encouraging to see the application of GRACE derived data particularly in Southern Africa given that GRACE is underrepresented. I would like to find out what was the time mean that was used to estimate the total water storage changes?
Tales: To give you further information about the GRACE satellite. The GRACE data have to be considered very careful. Why? Because it provides an overview on a large scale and not at local level. GRACE mainly ‘sees’ the fluctuations of shallow aquifers but not deep aquifers. Most of the time this is enough because most of the groundwater abstraction occurs from shallow aquifers. GRACE has been out in the skies since 2002. With the model we were able to add 20 more years going back to 1980 - which is fantastic. So this is what we tried to do, go back to the past to reconstruct these fluctuations.
Q: The applicability of GRACE on large aquifers (such as the Stampriet) you mentioned. So how is the applicability of GRACE on the smaller aquifers; and also the applicability of GRACE on a national scale for instance if you want to do it for the whole country total water storage estimations and what is the impacts on the water storage?
Tales: One of the aims of this study was to have a first picture of the correlations between groundwater and climate change at a large scale because there have been very few studies on that particularly in Africa. So, I would say a first step is to have a general big picture at large scale (i.e. the same study can be replicated at Southern Africa level) and then once we have that first picture, we can already identify some correlations. In Southern Africa, we can see that there is already an El Nino correlation and then you go into detail and further studies can be done. So, I can say that this study can easily be replicated at a large scale in other regions/ country level but is just to give a first ‘snapshot’ of the situation.. Then you follow with further studies at local level.
Q: Is there any reason why you chose this study area?
Tales: Well, we have chosen this area because we were working with the support of UNESCO in that region. We had support to assess these impacts of climate variability in the Stampriet aquifer and then we said ok let us see what is also happening in the Karoo Sedimentary aquifers so as to have a broad picture of groundwater dynamics in the Orange-Sengu Basin. We have also applied the same methodology of other regions of the world and the results were quite interesting.
Q: Did you fit groundwater level data with the modelled results? If so, how good was the model fit in validation?
Tales: If you have a look to slide 17. The answer is yes. We validated the model both in the GRACE (starting from 2002 until now) and the groundwater level data time-frame. Again, the groundwater level data might not have been representative of all of the basin but the very few that we could collect and were available, fitted very well and we felt confident to go further with these correlation of climate indices.
Q: Kwazi spoke about the scale with regards to applying GRACE data to a small scale. I personally tried that and the results were very coarse but looking at other studies, myself and others actually found that at scales greater than 150,000 km2 that is when you can start to see and get better results. We actually did not find absolute values in terms of values of total water storage but rather just the anomalies. Also using the soil moisture data and abstracting it from the GRACE the total storage data that actually gave us some insight into what is happening to groundwater level. So far we have not found or could use any long term data that is representative of areas less than the 150,000 km2 in SA and I think it is also the same issues that you faced when doing your study in the Stampriet. So I am not sure if you or somebody in your team have found some way to downscale so that can maybe be able to apply GRACE at a more local scale?
Tales: Yes, unfortunately for the time being we have not been able to go deeper into local scale. Exactly one of the challenging issues are the ones that you mention. So as you know groundwater fluctuates differently from one borehole to another, so a borehole that is only a few km away can have a different dynamic than another, and what we would need -I would say- to really make sure this model is robust is to have a strong groundwater level monitoring network at work in which you would allow you to have the average of all the groundwater level data. What GRACE does offer is usually an average of what is happening in the 150,000 km2 to give you one number. So that is the challenging issue. We did apply the same model in other aquifers that have a very good monitoring network mainly in the Unites States, and it worked very well. But as I said – this is only to show you what can happen at large scale.
Q: Is it possible to simulate the longer time effects of the Milankovitch cycles (100,000 year cycles) which effects of changes in Earth’s position relative to the Sun and are a strong driver of Earth’s long-term climate, and are responsible for triggering the beginning and end of glaciation periods (Ice Ages)
Tales: This is quite challenging as we don’t have good and reliable data on rainfall and evapotranspiration at long term to extend the model. The current version of the model starts at 1980 because our data on evapotranspiration starts by then.
Q: How long after the El Nino/La Nina has started do you start to see the impacts on the groundwater storage changes or was it not part of the study?
Tales: We could see here that it was highly responsive so only a couple of months afterwards we could already see the impacts. Again this is only for shallow aquifers but this is very important information in the sense that it is highly responsive. So if you know there is a drought coming you will have an impact very soon after and this is the reason why it is really important to really be in touch and discuss with the climate people e.g. the different meteorological agencies and so on, because if you can have a good forecast of what happens with the rainfall patterns then you can have a good forecast also for groundwater. In this case we have seen that the aquifers are highly responsive to rainfall patterns which are intrinsically linked with climate indices.
Q: What is your depth of "shallow" aquifers?
Tales: By shallow aquifers we mean here unconfined aquifer and the water table level that we could get here from the different chronicles were usually a couple of meters (approx. 10-20).
Q: What was the effect of human abstractions? How were these included in the model?
Tales: This is an interesting question as yes, it is difficult to get data on abstraction. This model here did not consider human abstractions but what we could see here and what we could see in other studies is that if you have a dry period it means that abstraction increases and the trend decreases. So in our study we talk about trends not numbers. The results of the model are shown in a normalized scale so abstraction can be “implicitly” considered because of that. It is very difficult to give you numbers on abstraction because it is not always very reliable data, so keep in mind the trends. The takeaway message here is that usually when you have a decreasing trend it also mean that you have more abstraction this means that the trend goes even down.
Q: The groundwater system seems to be more sensitive to human activities than climatic changes. Is long term still relevant or urgent? Long-term? 200 years or so
Tales: It is very difficult to separate what is the human component and what is the climatic component because both are together. When it comes to groundwater – if you have a drought – this means that you will abstract more so both are intrinsically linked. That is why it is very difficult to disaggregate both of them. In the long term, this is still relevant, because the information can help us to better set up some MAR schemes. For instance, we could see a AMO positive phase (current one since mid 1990s) could bring water storage down, and then once this cycle could flip to another phase we could have some better days. So this can already give us some time of what can happen and how we can prepare in the long term.
Q: Presumably the water level measurements were taken from boreholes, which normally would be associated with human abstraction?
Tales: Yes. We tried to collect the longest and most continuous water level chronicles. Unfortunately, there are very few.
Please contact Mr Kwazi Majola (Branch Chair GAU) for more information on this subject: MajolaK@nulldws.gov.za
SACNASP CPD EVENT
Well done and Thank you to the GWD Northern Branch for this excellent presentation by Dr Rainier Dennis!
Public domain borehole information in South Africa is generally stored in the National Groundwater Archive (NGA) and the Groundwater Resources Information Project (GRIP) databases of which both are centralized databases. The GRIP database is updated by the Department of Water and Sanitation, but only covers one of the nine provinces. The NGA on the other hand covers the entire country, however there is a backlog of borehole information that needs to be captured, and it has limited time series data.
The reason for the poor time series data is two-fold:
(i) groundwater monitoring is expensive due to the distributed nature of the resource (the NGA consists roughly of 280,000 boreholes over 1,225,986 km2) and
(ii) consultants tend not to upload data to the national databases as the data is seen as a competitive advantage.
During the recent drought experienced in the Western Cape Province (2015 to 2018), citizens of local communities took to social media, reporting on rainfall and groundwater levels within their communities. With the dams drying up people started targeting groundwater with the result of approximately 30,000 boreholes being drilled. This led to the development of a mobile app available to both citizens and groundwater professionals. This app allows logging of borehole information via smart phones. One of the main challenges with populating databases is the verification of the data.
The mobile app introduces a type of block chain approach where all data is accepted, but marked as low confidence until verified by a trusted user. The vision for the app is a ‘live’ hydrocensus and even if only water levels are captured, it would improve groundwater management by applying data mining techniques for trend analysis.
ABOUT THE PRESENTER
Rainier Dennis (BSc IT, BEng Electric & Electronic, MSc Geohydrology, PhD Geohydrology) is a senior lecturer at the Centre for Water Sciences and Management, North-West University. He has more than 18 years’ experience in software development, hydrological and geohydrological investigations.
We appreciate the sharing of the materials and we trust that Dr Rainier will receive a lot of additional inquiries around this inspiring development.
Recording available (please request with email@example.com)
Ivo: Is the code open source? if so pointers please
Once the App is complete, the code will be handed to the WRC and eventually to the custodian of the final database to make sure that if there are changes or future development, they can work with the existing code. So yes, it will be open.
Erin: What is the data consumption like?
Quite a difficult one. I think the best way is what Rob Schapers have said and that is he will start to track data usage as soon as the pilot version goes live. If you are worried about consumption the best thing to do is, at the office, download the data for that specific area so that the base map tile will be cache to your phone together with the borehole data that is available for that area. You can then go offline with your device and your GPS positioning device will still work in the field. You can log all the information and once you get back to the office you can connect to your Wi-Fi or any device you use and it will be uploaded. We will do some benchmarks in terms of just normal run-of-the-mill operations.
Irene: Two questions: (1) Has the feasibility of linking this to water use licenses/registration been considered to populate the database. And (2), will there be a function to convert what3word coordinates to surveyed data for use by groundwater professionals?
The first question concerning water use licenses, no, we haven't considered this. My guess would be if people would be using it they may see the applicability and we might get some more requests to add this type of functionality - it is a good idea. The second question, yes on the use of the what3words. It can convert either way – by the coordinate of the block or the what3words can take you to the center of the block but remember, since this is a GIS system all those boreholes that you see on your display already has its own latitude longitude built-in so, it is not really necessary to convert back from the what3words as the GIS already have a coordinate for each site.
Nico: Is there some security measures in place to prevent the public to be targeted, due to location and photo sharing. Some chap might be interested in high-jacking your pump and sell on the black-market, if your borehole is not properly secured.
Yes, obviously there are certain things and this is a very valid point. We’re also sitting with the POPI – the Protection of Personal Information Act so you can't publish owner’s details and so forth. That is something that we obviously need to take cognisance off and being aware of this for some time, we don't have a solution yet. You can obviously strip the GPS and location details from any image but if you have the phone and you’ve registered on the App., obviously you can see there is a borehole and it’s got a pump in it, so, yes that might be a concern.
Wilbè: Wouldn’t it be better to develop the apps natively? If the source code is available I could assist in writing a MAC desktop version if there are any mac users
At the end of the day what the App studio does, is when the QML (Qt Markup Language) is compiled it invokes the relevant c++ compiler for each target platform, so in essence what you end up on your phone is a native app. The only reason why we have used the Arc products is because we are making use of ArcGIS online and you can also run some analysis of the data from the ArcGIS online server and then eventually make that analysis also available to users. If you were to develop that from scratch it would be a massive undertaking.
Jorette: Great project! On some projects we used to overlay the NGA, WARMS and whatever data we could over Google Earth to assist with hydrocensus investigation in the field, but phone battery dies and doesn't always show your position in relation to the nearby boreholes. Have you had similar problems or lags?
I suppose this will depend on your specific device. In our testing we have not come across this, but to date we were only using an iPhoneX and a LG V20 in our testing. When you disable the data network on our app, assuming you have cached all the data for your study area, you will conserve power and lengthen your battery life. I am not sure if Google Earth provide the same functionality to save battery time. It is also worth mentioning that GPS accuracy on a mobile phone, only making use of GPS (no Wifi) out in the open can range from 1m to 5m. This accuracy can be increased if making use of external GPS antennas, but from a citizen science point of view we don’t expect users to do this. Some phones support high accuracy mode which is switched on in the Settings menu.
Please contact Dr Rainier Dennis or Prof Ingrid Dennis for more information on this subject: Prof Ingrid Dennis <firstname.lastname@example.org>, Rainier Dennis <Rainier.Dennis@nullnwu.ac.za>