GRACE for total water storage estimation

GRACE is a remote sensing of water storage based on gravity field. The two twin NASA's Gravity Recovery and Climate Experiment (GRACE) mission detect the gravitational field of the Earth surface. Since the effect of land mass (or large body) in the change of gravitational field is small, and slow, the main change is due to the change in water mass. The terrestrial water has a mass, and always in motion. So measuring the time varying gravitational field is measuring the total water storage of the earth. This gives the hydrologist unprecedented chance to see and understand how much they are able to close the water budget modelling.  Since  the total water storage (TWS) is the main sources of water for use, and it the aggregated values of the water fluxes, the GRACE data will help the water resource managers and policy planners  to be very proactive.  

Recently, I have been starting to explore this dataset if it is possible to use in basin hydrological modelling. While I will come back to the results of the study in the future, here is the GRACE map of Ethiopia for some years (2002-2008) at monthly time steps. I presented the map in animation below. It is estimated at zonal level. Note that the unit is cm.   

"Poverty reduction in Ethiopia in the last decade have been closely associated with unusually reliable rainfall"

Last week  in Addis Ababa EGU conference, I met Professor Simon Dadson who is the co-investigator of this new and large (from 2015-2022)  project called REACH . The project is funded by the UK government and hosted by oxford university school of Geography and the Environment, aiming to conduct water security and poverty issues in three countries:  Kenya, Ethiopia and Bangladesh.   This is really the beginning of  interesting effort, and I hope  it will accrue wealth of  knowledge and understanding on the clear  relationship between water resources availability (security) and level of poverty. 

While I will follow their efforts and works closely in this regards, at the moment, browsing  the project website I have seen this interesting statement "Major reductions in poverty [in Ethiopia] in the last decade have been closely associated with unusually reliable rainfall".  It is true that small land-hold rain-fed agriculture farmer in Ethiopia depends mainly on nature of rainfall. However, I never noticed this contribution to the recent 'poverty reduction' process. In fact the climatological drought we have this year due to El Nino is  highly (directly) contributing to agricultural drought and famine.  

Understanding the rainfall patterns, or be able to forecast it, and estimation of  its effect on hydrological and agricultural drought is crucial step. If possible,  hydrological forecasting for future short time such as for one/two week(s) would really be very interesting challenge that I would like to do in my academic career life!

EGU Topical conference

This year, three important meetings of water science  jointly organised at Addis Ababa, from November 18-20, 2015.  The three meetings are the Alexander von Humboldt Conference of the European Geosciences Union, the STAHY workshop of the International Commission on Statistical Hydrology of the International Association of Hydrological Sciences (ICSH-IAHS) and the Leonardo Conference of the Hydrological Sciences Division of the European Geosciences Union. While there were a lot of good presentation, the following three presentations were more interesting topics of  basin water balance modelling  in general and our (my) approach of doing water balance estimation in particular. 

• Towards Optimization of Reservoir Operations for Hydropower Production in East Africa: Seasonal Climate Forecasts (Leonardo Lecture) --- by Mekonnen Gebremichael. Abstract 
• Education and TAHMO, the Trans-African Hydro-Meteorological Observatory --- by Nick van de Giesen. Abstract  
• How important are soils for hydrological modelling?  --- by Hubert H.G. Savenije. Abstrac 
• characterisation of the regional variability of seasonal water balances within the Omo-Gibe River basin --- by Adanech Yared Jillo 

I had two presentations: One on the comparison of satellite rainfall estimation products for the purpose of basin water balance modelling inputs, and the second was a work in progress on JGrass-NewAge set-up for water balance estimation in Upper Blue Nile basin.  The abstract can be found here and here, respectively. A video of  one of the my presentation (or a part of it) is recorded by my friend and can also be found in this you tube. 

List of Hydrologist working in Ethiopia and horn of Africa

Hydrologists all over the world, I mean those who are real and good ones, are peoples fighting with nature to understand better how nature (hydrological cycle  works, and are those providing and accumulating hydrological information. So having any good hydrologist at any corner of the world is obviously mean better awareness and insight how the terrestrial water cycle. 

However, when it comes to solve a particular water related problems in a particular area, those who did researches in that regions are better understand the problems and provides "quantitative" information. More over, every basin is hydrologically unique! 

Ethiopia is hydrologically complex country with many water related problems. Having a personal goal to deal with the water related problems and particularly agricultural water resources in Ethiopia, I am interested to talk, to discuss and collaborate, or even to follow and read their research outputs. So here is an updating list of hydrologist that works in Ethiopia and surrounding regions. The list will be those of pure hydrologist, in the sense that I will not include other related scientists. And I don't pretend that I will find all hydrologist working in that regions. It can also be used as list to organise workshop and conference in the region.  Anyway some of hydrologist working in the region are:

• Assefa M. Melesse (GS)
• Declan Conway (GS)
• ALEMSEGED TAMIRU HAILE (IWMI)
• Simon Langan (RG)
• Tammo S Steenhuis (CU)
• Shimelis Gebriye Setegn (GS)
• Seifu A Tilahun (GS)
• Senay Gabriel (SDSU)
• Paul J. Block (CV)
• Tom Rientjes(GS)
• Mekonnen Gebremichael (UCONN)
• Seifu Kebede (RG)
• Tenalem Ayenew (RG)
• Dagnachew Legesse (RG)
• Amy S. Collick (RG)
• Semu Moges (RG)
• Tena Alamirew (linkedin)
• Meron Teferi Taye (GS)
• Menberu Meles Bitew (GS)
• 
  

To be updated …….

Short Announcement: SRTM 1 Data Worldwide

Short Announcement: SRTM 1 Data Worldwide

satellite rainfall estimation products in Upper Blue Nile Basin

This is stub for the complimentary material for our paper: Comparative evaluation of  satellite rainfall estimation products and bias correction in Upper Blue Nile Basin. I will provide the data, R scripts and some supplementary results soon....

• The Satellite rainfall dataset 
• The ground-gauge dataset 
• The DEM 
• R scripts 

All will be filled soon


Table: Bias correction function for different SREs in UBN basin

The small, the large, and the macro hydrological system

The global hydrological cycle is a closed system, meaning that the amount of water is fixed.  No input, no output. It is just the circulation! Every river basin  takes the amount of water it needs for its ecosystem maintenance and return it back to the global hydrological system. For a river system, the input to the  basin is precipitation whereas the outputs are  the amount of water that the basin return it back to the system. These are the discharge ( river flow), and  evapotranspiration driven by the energy balance.  As the system changed from the closed system to the open system as moving to the global cycle to basin water cycle,  the basin scale at which the hydrological cycle is looked at  matters. This mean that the proportion of the components such as the precipitation, discharge, evapotranspiration, storage varies across scale. 

The first procedure in modelling the hydrological system is the geometry at which the cycle is estimated. This geometry is extracted from the digital elevation models. These days, they can be easily available from different sources.    Based on the objective and purpose of the modelling, the spatial scale of the basin  can be ranged from few kilometres to hundreds of kilometres (or continental scale).  The digital watershed modelling (DWM) is the pre requisite for modelling, for instance, we are interested at different scales, and  we have been working in the following:

1. Posina Bain, small scale basin

1. Adige scale, large scale basin 

1. Upper Blue Nile basin, Macro scale 

2. Adige river basin

I will not talk about Posina basin in this post. I will have other post about small basin DWM, and hydrological modelling  space-time variability.  Adige is one of the largest,   the second largest basin in Italy(?).  It provides water resources to all the Bolzano, Trentino and Veneto region. We have interest to  model water resource at this basin, and the first step is  the DWM. It is possible to start from the whole basin, and look it into the detail.  To work on the maximum detail topographic information, following series of steps as described in other post, DEM need to be partitioned into many detail, for instance, here the Adige is divided into about 1200 HRUs.  

Adige basin partion into 1200 HRUs using JGrass Spatial toolbox

However, such large numbers of HRUs could computationally be demanding and difficult for data management,  particularly if we are interested to the hydrological outputs at each HRUs. For this reason, the basin can be separated into major basin, and the simulation can be take care of at each particular basin, and use some routing system to estimate at the furthest outlet of the whole basin.  For instance, Adige basin can be divided into several basin (notice the black divides inside the basin, in to five major  basin), and then hydrological simulation can be carried out at each basin, and some sort of routing mechanism can be applied to route to the outlet. 

Some of these can be: 

The position of Adige-Passirio basin (right)  and the topography partitioning into HRU 

 

            Isarco  basin (relatively small basin) that can be singled out for simulation purpose

                    Rienza basin and its topographic partitioning 

Avisio basin and its partition

Noce basin and its partition

3. Upper Blue Nile basin

What we have to do if we are interested even larger (very larger ) basin than the Adige ??? For instance Upper Blue Nile basin, the Ethiopia part of the Blue Nile?  Let's start from what people already did: 

   

 Lake Tana basin where most hydrological studies in the UBN basin is conducted 

• e.g Alemseged T. Haile, Tom Rientjes, Ambro Gieske, and Mekonnen Gebremichael, 2009: Rainfall Variability over Mountainous and Adjacent Lake Areas: The Case of Lake Tana Basin at the Source of the Blue Nile River. J. Appl. Meteor. Climatol., 48, 1696–1717. doi: http://dx.doi.org/10.1175/2009JAMC2092.

Senay, G. B., Asante, K. and Artan, G. (2009), Water balance dynamics in the Nile Basin. Hydrol. Process., 23: 3675–3681. doi: 10.1002/hyp.7364

At this study of water balance at the whole Nile basin, UBN is used as single basin. No space-time variabilities study 

Dile et al, also studied water balance at Lake Tana basin ..... as example for UBN basin study 

 Dile, Yihun Taddele and Raghavan Srinivasan, 2014. Evaluation of CFSR Climate Data for Hydrologic Prediction in Data-Scarce Watersheds: An Application in the Blue Nile River Basin. Journal of the American Water Resources Association  (JAWRA) 1-16. DOI: 10.1111/jawr.12182

Gebrehiwot et al, selected 12 basins a way of looking spatial variabilities in water balance  temporal dynamics  

Gebrehiwot, S., Gardenas, A., Bewket, W., Seibert, J., Ilstedt, U. et al. (2014). The long-term hydrology of East Africa's water tower: statistical change detection in the watersheds of the Abbay Basin.

Similarly for looking space-time varibaility, Tekleab et al  select 18 basin and modelling the water balance at those sub basin. 

Tekleab, S., Uhlenbrook, S., Mohamed, Y., Savenije, H. H. G., Temesgen, M., and Wenninger, J.: Water balance modeling of Upper Blue Nile catchments using a top-down approach, Hydrol. Earth Syst. Sci., 15, 2179-2193, doi:10.5194/hess-15-2179-2011, 2011.

"  .......a conceptual process-based hydrological model is developed .....to study the runoff mechanisms and rainfall–runoff processes..."

Dessie, M., Verhoest, N. E. C., Pauwels, V. R. N., Admasu, T., Poesen, J., Adgo, E., Deckers, J., and Nyssen, J.: Analyzing runoff processes through conceptual hydrological modeling in the Upper Blue Nile Basin, Ethiopia, Hydrol. Earth Syst. Sci., 18, 5149-5167, doi:10.5194/hess-18-5149-2014, 2014.

  

ANIL MISHRA & TAKESHI HATA (2006) A grid-based runoff generation and flow routing model for the Upper Blue Nile basin, Hydrological Sciences Journal, 51:2, 191-206, DOI: 10.1623/hysj.51.2.191

This study is grid-based distributed analysis, however, the focus is on runoff production, where evapotranspiration and other component are not explicitly considered.

 Kim and Jagath made interesting set-up to  model the UBN using 6 sub-basin! However, space-time  modelling of UBN using 6 is not something to envisaged ...... 

Kim, Ungtae and Jagath J. Kaluarachchi, 2009. Climate Change Impacts on Water Resources in the Upper Blue Nile River Basin, Ethiopia. Journal of the American Water Resources Association  (JAWRA) 45(6):1361-1378. DOI: 10.1111 ⁄  j.1752-1688.2009.00369.x

Obviously, further space-time hydrological budget characterisation is necessary in UBN basin. In contrast to the maximum subbasin of 6  partition used to study the whole UBN basin, we need to decide some level of partition that really capture the true variability of the rainfall, evapotranspiration, runoff and storage function. 10 ? 20 ? 50 ? 100? 200? 500?  For the moment, lets try with 300 HRUs. 

Upper Blue Nile basin discretisation into 300 HRUs

To be continued........

2015 World Water Week summary

There is very interesting international annual conferences called World Water Week. It is organised by Stockholm International Water Institute (SIWA) to discuss the globe’s water issues. This year, the theme is Water for Development.  Some of the presentations and plenaries are available here at this SIWA media hub, and I find it very informatics to explore the status of water resources globally. Since the talks are very diversified, you need to explore to the presentation you are interested in! 

(Ensemble) Hydrological Modelling, calibration, and FUSE

Claudia Vitolo has a series of R tutorials for using a single hydrological model, or a combination of models to estimate hydrological component using a modular framework to diagnose differences between hydrological models.  I find it very interesting, and seems the procedure can be integrated to be used in our approach of estimating large scale hydrological components using JGrass-NewAGE system. That mean while the FUSE can be used to estimate the any component at the subbasin scale and use to estimate the uncertainty in the prediction,  while the  JGrass-NewAGE system can be used to integrate the process to estimate at large scale. This is in my list to try in near future! 

The three tutorial are:

1.  FUSE model in RHydro package (part 1: simple simulation) Here
2. FUSE model in RHydro package (part 2: calibration) Here
3. FUSE model in RHydro package (part 3: ensemble) Here

Enjoy it ! 

Water stress country ranking

The WRI published a report on  the level of water extraction in relation to the available surface water, and able to come up with water stress  ranking of all countries. Water use in different sector such as domestic consumption, agricultural farms, and industrial sectors will be influenced by the change in the water availability, and will have direct impact on economic development.   Anyway, the 10 most stressed countries are  Bahrain, Kuwait, Palestine, Qatar, United Arab Emirates, Israel, Saudi Arabia, Oman and  Lebanon. The whole report can be read at here. 

This map shows  spatial distribution of the  projected  water stress by 2040: some of the high projected stress countries are middle east countries, India, China, South Africa, Botswana, Australia, Morocco and Tunisia, Libya, USA, etc

Drought in Ethiopia

According to different sources such as relief web , news24, and European commission for humanitarianism aid,  in this year, there has been a drought in Ethiopia, particularly in the Easter and Southeastern part of the  country.  I have been closely following the news in this regards, and some of the News  shows the level of severity on human is  really shocking. I am  deeply sadden.  For a glimpse view  on the nature of the drought is documented by this aljazeera video.  According to the European commission for humanitarianism aid  Easter Tigray, Wollo, Shewa, Arsi, Harerghe and Somalia areas are at risk. FEWS (http://www.fews.net/east-africa/ethiopia) presented the spatial distribution of the famine food insecurity of the few  previous months and short term forecast estimation is shown in the following map.

The first thing is first! We need to help those people at risk. We need to feed those poor peasants suffered by this unpleasant evil, famine, and save their  life. Most of the time, Ethiopian farmers are very dogmatic and religious to such natural disasters, and  they mask their suffers and agonies.  By the time  the drought is discovered, it could be already late and the risk on human life could be very high. Worse than that, most of the time, all governments of Ethiopia has been always reckless about the situation , and they only worries about their power and refrain from accepting the fact. There,  it takes another time to nationally accept the problem and act for the solution. IT is ok to play the political game, but HUMAN FIRST please!
Secondly, if the rainfall is short to provide the peoples' demand, we have to look for other option. For instance,  for those pastoral areas, the government could have  invest to drill some walls and supply water for animals. For pastoralist,  few sites of water walls could serve large communities.

Fourthly, if not possible to maintain most the incoming rainfall into the blue water system, the next strategy is to make an effort to infiltrate the rainfall into the soil and store it in the form of what is called green water. So all the land management activities such as terrace and conservation tillages helps to infiltrate the rainfall.  In a region where there is high potential evapotranspiration, the water will leave back to the atmosphere. If we don't use this stored water, it will leave as evaporation which is from bare land and which is non-productive.  However, all the idea of the use of green water means that the return has to be through the plants in the form of transpiration, after producing the biomass instead of just evaporation from the bar lands. Hence, our land has to be covered in all the vegetations and increase the availability of the green water.  In other words, our technology has and understanding of this transpiration need to clearly outline important crops (local climate adapted species ) so as to increase the water efficiency.    SHIFT the transfer (cycle) from evaporative loss to useful transpiration!
Lastly, there has to be forecasting and prediction of those events. We know drought (famine) more than any country in this whole world. On the contrary, we are the least to develop the strategies and methods to  forecast (predict). We don't have to say "the cause is El Nino effect" for every year's drought. The fact that the drought is happening almost in every two years specifically in this region, this reason does not convinced me at all. We really need to work what is exactly happening in the regional climate circulation and,  be able to forecast  the subsequent effect on the hydrological  cycle of the region!   


Huge number of animal is dying in some part of Afar and Somalia regions, and hence, life of the people  depends on animal, pastoralist, are at high risks. Now the drought extends to the small land holding farmers in the above mentioned area.  As reported from the News that the cause of the drought is  rainfall shortage. Drought is essentially natural, and happens in different part of the world. It is unavoidable!  However, how we react to the drought is what matters and what makes difference. The point is what has to be done and what strategies has to be designed to reduce the impacts of drought?

Thirdly, in semi-arid region, the focus has to be on water productivity, not on land productivity. The first step is to make an effort that the income water, rain, into the blue water system (small pond and natural and artificial lake). If we have water storage that we can supplement the rainfed agriculture, then, for sure, with all the degraded soil productivity, the farmers will have enough food production. INCREASE the accessibility of  BLUE WATER for the farmers so as to increase  water productivity! Blessed are those invest in the irrigation projects!

Creating a research brand

In the search of approaches and tips claiming to be expert in a particular research theme , today, I have found that   Jeffrey J. McDonnell publish small advice on science. The advices can be summarised as:

• Define research theme or brand that interests you and that you can do it very well

•  Define proper research brand width: not too much wide and too much narrow 

• Disseminate it to the world everywhere you get chances: through published  paper, conference   talk, through research webpage,  journal commentary, perspective paper, or review linked to your research brand

You can read it for yourself here.

Similar advices has been published by my advisor, Riccardo Rigon, at his blog abouthydrology.  For instance, first one is some general comments how to do research in general and how to follow "smart" person (here), and the second is very specific to hydrologist and some practical advices (here). 

The overall message is that its is very important for one early researcher to define research agenda (realm) and follow persistently to standing out among the many young scientists. And follow those mentioned advices for improvements.  As to my research agenda, well, at the moment I will not have much agenda that is not agenda of my  Professor, but, I would like to connect those research brands that interest me and I am very good at it. For the moment   they are described in this post, and will be further refined.  

Hydro meteorological data bank of my blog

To answer our research question and hypothesis, WE hydrologists use different kinds of data both  hydrometer (e.g discharge amount and its properties, channel hydraulic information, sediment and its properties) and meteorological data (such as rainfall, snowfall, temperature air wind, humidity, air pressure and solar radiation etc). It  comes from different sources. It can be from ground measurement, remote sensing instruments. Once we used them to analyse specific research issue,  it important to store

for future use.  I will dedicate this post for data storages that comes from different sources. If someone is interested to use readily (at least downloaded and extracted at the upper Blue Nile basin)  available data, s/he can contact me.

Some data (to be updated)

1. TRMM daily rainfall data extracted from NASA for UBN basin can be find here 

1. TAMSAT daily rainfall data for UBN kindly provided by TAMSAT Research Group & NCEO Impact Team, University of Reading can be find here 

1. CMORPH will be here soon 

1. CFSR will be her soon 

1. SM2RAIN will be here sooon 

Quantifying groundwater or terrestrial water storage using GRACE

The use of GRACE data for hydrological models to quantify the terrestrial water balance is a way to approach the large body of unknown knowledge of the water storage function. While the use of models and meteorological data helps to estimate most of the water balance components with some reasonable confidence, the quantification of the storage is very difficult, and such supporting data is clearly improve the estimations particularly in large basins. Sooner or later, in the line of water balance estimation at large scale basin, the use of GRACE will be my research agenda. I came across to this paper which shows some approach how to include the GRACE data in the basin water balance modelling.   The abstract reads as:

"Groundwater is an increasingly important water supply source globally. Understanding the amount of groundwater used versus the volume available is crucial to evaluate future water availability. We present a groundwater stress assessment to quantify the relationship between groundwater use and availability in the world’s 37 largest aquifer systems. We quantify stress according to a ratio of groundwater use to availability, which we call the Renewable Groundwater Stress ratio. The impact of quantifying groundwater use based on nationally reported groundwater withdrawal statistics is compared to a novel approach to quantify use based on remote sensing observations from the Gravity Recovery and Climate Experiment (GRACE) satellite mission. Four characteristic stress regimes are defined: Overstressed, Variable Stress, Human-dominated Stress, and Unstressed. The regimes are a function of the sign of use (positive or negative) and the sign of groundwater availability, defined as mean annual recharge. The ability to mitigate and adapt to stressed conditions, where use exceeds sustainable water availability, is a function of economic capacity and land use patterns. Therefore, we qualitatively explore the relationship between stress and anthropogenic biomes. We find that estimates of groundwater stress based on withdrawal statistics are unable to capture the range of characteristic stress regimes, especially in regions dominated by sparsely populated biome types with limited cropland. GRACE-based estimates of use and stress can holistically quantify the impact of groundwater use on stress, resulting in both greater magnitudes of stress and more variability of stress between regions."

some references (to be updated ):

Van Dijk, A. I. J. M., L. J. Renzullo, and M. Rodell. "Use of GRACE terrestrial water storage retrievals to evaluate model estimates by the Australian water resources assessment system." Water Resour. Res 47 (2011): W11524.

Kirk Zmijewski and Richard Becker, 2014: Estimating the Effects of Anthropogenic Modification on Water Balance in the Aral Sea Watershed Using GRACE: 2003–12. Earth Interact., 18, 1–16. doi: http://dx.doi.org/10.1175/2013EI000537.1

Deus D, Gloaguen R, Krause P. Water Balance Modeling in a Semi-Arid Environment with Limited in situ Data Using Remote Sensing in Lake Manyara, East African Rift, Tanzania. Remote Sensing. 2013; 5(4):1651-1680.

van Dijk, A. I. J. M., et al. "A global water cycle reanalysis (2003–2012) merging satellite gravimetry and altimetry observations with a hydrological multi-model ensemble." (2014).

Eunjin Han, Wade T. Crow, Christopher R. Hain, and Martha C. Anderson, 2015: On the Use of a Water Balance to Evaluate Interannual Terrestrial ET Variability. J. Hydrometeor, 16, 1102–1108. doi: http://dx.doi.org/10.1175/JHM-D-14-0175.1

Scanlon, B. R., L. Longuevergne, and D. Long (2012), Ground referencing GRACE satellite estimates of groundwater storage changes in the California Central Valley, USA, Water Resour. Res., 48, W04520, doi:10.1029/2011WR011312.

AboutHydrology: R resources for Hydrologists

The R resources for hydrologist post is one of the richest site for hydrologist to review the available R packages out there.  You can find it here AboutHydrology: R resources for Hydrologists: it is very amazing collection of resources. Using the same title, I am dedicating this post to collecting all R packages and resources that I am using and come across.

mco: Multiple Criteria Optimization Algorithms and Related Functions. Example application 

gtools:smartbind: Efficient rbind of data frames, even if the column names don't matchspot: An R Package For Automatic and Interactive Tuning of Optimization Algorithms by Sequential Parameter Optimization

Displaying Time  Series, Spatial, and  Space-Time Data  with R: Interesting book for spatial and temporal data visualisation

RStoolbox: Provide a set of high-level remote sensing tools for various classification tasks. All list of function can be found here. 

The spatial and temporal evolution of contributing areas: Nippgen et'al, 2015

I am always interested on the representation of topography and their conditioning into the hydrological modelling. I have seen this accepted paper of Brian McGlynn and his co-workers entitled "The spatial and temporal evolution of contributing area". At the moment, I didn't go through it  but the topic is very  interesting, so I am saving to my list to read papers. The introduction section (objective of the paper ) reads as follows:

"Here we present a parsimonious but fully distributed modeling framework that incorporates topographically driven lateral water redistribution and eddy covariance derived spatially disaggregated evapotranspiration measurements to simulate streamflow and the spatial distribution of water stored in the watershed through time. The goal of the model development and this particular application was to inform our understanding of the spatial and temporal dynamics of runoff source areas. Utilizing and testing this diagnostic modeling framework in conjunction with empirical measurements of hillslope connectivity through an extensive network of shallow groundwater wells [Jencso et al., 2009; 2010; Jencso and McGlynn, 2011] we approximated watershed connectivity to investigate how runoff source areas change in space and time over the course of two water years in a snowmelt dominated system."