Modeling_Nov_2012

=Symposium on Modeling in the Blue Nile / Abay Basin=
 * 12 November 2012, ILRI Addis Campus**

toc

Objectives

 * Mapping out and sharing modeling experiences in various projects.
 * Identifying priorities and data needs for ag-water and water resource modeling in the Blue Nile Basin.

Agenda

 * **8.30 Registration**
 * //**9.15 Opening / welcome**//
 * **Participant Introductions**
 * **Meeting overview**
 * **Partner presentations / Modeling for Planning:**
 * [[file:Deksyos Preentaion_Tana_Beles_Nov2012_IWMI.ppt|NBI National Nile Basin Decision Support System Office]] - Deksyos Tarekegn
 * Abay Basin Authority - Dejene Sahilu /Yewondwosen Mengistu
 * [[file:Anteneh_Dejenie_IWMI_NBDC_Symposium.pdf|Tana Sub-basin Organization]] (TaSbO) - Birlew Abebe /Anteneh Zewdie
 * [[file:Modeling_moges final.ppt|Beles Sub-basin Organization]] (BeSbO)- Moges Berbero /Alemayehu
 * [[file:modelling_experience_MoWE.ppt|Ministry of Water and Energy]] - Semunesh Golla /Abity Getaneh
 * [[file:presentation_IWMI_Seifu.ppt|Barhidar University]] - Seifu Admassu Tilahun
 * [[file:adanechYared.pps|Ethiopian Institute of Water Resources]] (EIWR) - Adanech Yared, PhD candidate
 * //10.30 Coffee//
 * **Partner presentations continued…**
 * **NBDC modeling overview** (depending on timing)
 * //12.00-1.00 Lunch//
 * **NBDC Modeling Overview**:
 * [[file:Augmenting_Hydro-MET_Data_Demands.ppt|Augmenting hydro-MET data demands of biophysical models]] – Solomon Seyoum/ Charlotte MacAlister
 * [[file:presentation_IWMI_Seifu and tammo.ppt|Hydrological process modeling]] – Solomon Seyoum / Tammo Steenhuis
 * [[file:Integrated_Modeling_RMS_Impacts.ppt|Water resource modeling]] – Solomon Seyoum/ Charlotte MacAlister
 * [[file:teklu-Crop Water Productivity Modeling.ppt|Crop water productivity modeling, demand and impact at a field level]] – Teklu Erkossa
 * [[file:Addisu_ambo.ppt|Integrating farmers practices and perception into crop modeling]] – Addisu Asfar, MSc Candidate
 * [[file:Presentation N4 Modeling Workshop (Nov 12_2012) Kindie G.ppt|Socioeconomic impact assessment at a landscape level]] – Kinde Getnet
 * Rainwater Management Targeting tool - Charlotte MacAlister
 * [[file:RITZEMA_BN_Mod_Symposium_20121112.ppt|Integrating biophysical and socioeconomic model outputs for MCA]] – Randall Ritzema
 * //2.45 Coffee//
 * **Roundtable Discussion**: Identifying priorities for ag-water and water resource modeling in the Blue Nile Basin.
 * Scaling issues – integrating small scale practices and large scale impacts in planning and management
 * Data needs
 * Accuracy and uncertainty
 * Uptake and acceptance of model outputs
 * Application of models at a household / community level
 * Plenary - Feedback
 * Closing

NBI decision support system office (D. Tarekegn)
Huge irrigation potential for both sub-basins (Tana and Beles). MIKE BASIN is a simulation model for water allocation representing the hydrology of the basin in space and time.

Tana Sub-Basin Organization (Y. Mengistu)
The TaSBO uses models for evaluation and improvement of the reliability of hydrological data (fill gaps in previous data, predict future conditions) and for simulation and scenario analysis (strategic development planning). Challenges: massive intervention in land and water resource use (change in water use behavior, climate change etc.), limited number of modelers, limitation in the tools --> required system thinking and integration of models. Using the **Tana Beles DSS** for flow measurement etc., **SWAT** for generating stream flow data (for prediction of water flow), **WEAP/Mike Basin** for sector demand analysis, **MODFLOW** (under construction) as a conceptual model for flow from the basin to the Beles basin, **GIS-RS** modeling (e.g. for vegetation dynamics).

What will be the impact of upland soil and water conservation practice? Will it enhance recharge?
 * Challenges**: Selection of better modeling tools, capacity of professionals, consistency of use with selected models, regular updating of models etc.

__Comments__:
 * Land use is a major issue. Before human interference it was all ok. It's important to plant trees to anchor the roots of plants to avoid sliding. It increases the stability of the soil and the slope. When storing water, the physical structures depend on soil and slope. If the latter is not stable, it's not advisable to keep the water there. We have to drain the water from the slope without causing erosion. Drainage is very important.
 * We don't know the solution at this point. We have to investigate this further.

Beles Sub-Basin Organization (M. Berbero)
BeSBO working on reearch and special studies. Using models to assess surface and groundwater in the basin (using SWAT and MODFLOW models), water resource allocation and planning in the basin (using WEAP and MIKE BASIN).


 * Challenges**: existing meteorological stations are not uniformly distributed. Existing hydrometric flow measurements did not cover all the catchment area of Beles basin etc...

__Q&A__:
 * **Q**: Where did you get your soil information?
 * **A**: Soil database discussed with Solomon - provided by IWMI.
 * **Q**: What are the major initiatives to improve the quality of station data, soil maps, flow data etc.? Is it acknowledged as a problem? Do you work with AFSIS? There is also EthioSIS. ICRAF is heading this work in Nairobi but in Ethiopia, ATA is taking care of EthioSIS.
 * **A**: We recognize the problem but don't know about these initiatives.

Ministry of Water and Energy (S. Golla)
Use of models to inform decisions about quantity and quality of water resources, and water management in complex regional systems. Models: simulation and optimization, hydeological & meteorological, disaster management, water quality, design models. MoWE has been developing models as part of the Integrated Basin Development Master Plan Studies. Most models developed by consultants - some by in-house staff. Models used: **HEC-5**, **WATBAL**, **RIBASIM**, **MIKEBASIN**, **WEAP**, **SWAT**, **Mike 11** etc.


 * Capacity and challenges**: Very low capacity. Some practical experiences but generally need to establish modeling unit. Challenges include: no adequate experienced staff, high staff turnover, lack of modeling tools and equipments. No take over of the activities undertaken by different studies.
 * Future direction**: Check, update and make use of existing models; integrate among different directorates; capacity development (human and institutional), cooperate with different stakeholders, universities, institutes working in modeling.

__Q&A__:
 * **Q**: Why do you use different models for master plan management?
 * **A**: Different initiatives were organized at different times with different consultants who used different models. We need to harmonize models.
 * IWMI would be happy to help in this. A lot of organizations are not keen on sharing their models but bringing initiatives together would speed up the use of modeling throughout. It is possible to use different models that are well calibrated and to bring together results without having to invest in only ONE big DSS.

Bahir Dar University presentation 'Storm runoff and soil erosion processes on the Ethiopian Highland' (S. Admassu Tilahun)
Modeling required to investigate where runoff & erosion processes take place in the landscape. PEER project (USAID-funded) - Partnerships for Enhanced Engagement in Research - proposes more effective soil and water conservation practices by identifying those parts in the landscape that contributes most of the sediment and nutrients at the outlet. Its activities: measure stream flow and sediment concentrations, measure shallow ground water level, employ sediment tracers to locate sediment source areas, focus group discussions and transect walk, soil nutrient analysis.

__Q&A__:
 * **Q**: Have you checked the volume of sediments? After August discharge is high.
 * **A**: We observe the same pattern: sediments decrease.

Ethiopia Institute of Water Resources (A. Yared)
Methods used in the project: Naturalizing stream flows, using **WEAP**, other primary & secondary analysis, flow modeling and impact analysis after changing the flow in downstream ecosystems etc. Looking into rainfall and runoff variability.


 * Challenges**: Limitation of data

Augmenting hydro-MET data demands of biophysical models (S. Seyoum)
Modeling landscape processes requires detailed climatic and geographic datasets. Meteo stations in most parts of Africa are very sparse. Cimate records are incomplete  High-resolution glob reanalysis data for SWAT modelling applications in Africa. Climate forecast and reanalysis system (CFSR): A coupled atmosphere-ocean-land-sea-ice system; accounts for changing CO2 and other traces gases. Finer spatial and temporal resolution. Applications of CFSR data: a) for weather generator files for areas with missing / incomplete climate datasets b) SWAT weather input files for un-gauged water sheds c) climate downscaling and bias correction d) study of large scale water and energy fluxes. CSFR for water fluxes study helps identify hot spots.


 * Conclusions**: high resolution reanalysis data has great potential to improve modelling of landscape processes.

Hydrological process modeling (T. Steenhuis)
Usually requires a lot of distributed parameters that are validated. More input parameters do not necessarily give better results. What about watershed management?

Water resource modeling (S. Seyoum)
Using SWAT model around climate, soil, LU, TI and WEAP around landscape development scenarios and verging onto impact evaluation. How to incorporate both models for basin planning?

WEAP modeling to analyse large scale developpment interventions. Improved schematization for SWAT-WEAP linkages. SEI provided first feedback on the WEAP setup. Scenario definition (development + RMS). SWAP modeling looking at land cover maps, farming system maps etc. Combined it provides a land-use map. Ending up with topography-soil map. Climate data: using CFSR data.


 * Q&A**:
 * **Q**: Benefits you are trying to maximize are transpiration and recharge to aquifer.
 * **A**: Even recharging aquifer helps maximize evaporation.
 * **Q**: Do you have the results?
 * **A**: We do, some but it's not published yet.

Crop water productivity modeling, demand and impact at a field level (T. Erkossa)
Lots of demands from downstream users, hydropower etc. --> we need to improve water productivity, which can be defined in various ways e.g. biophysical product, actual evapotranspiration, rainfall etc. Vertisol areas in the basin, divided in 2: drainable or non-drainable areas. Use of **CROPWAT**.


 * Impacts**: use of BBF increased water demand but reduced evaporation loss. BBF increased WP with respect to effective rainfall. Growing rice instead of grass-pea or grazing increased water demand but reduced evaporation. Use of BBF on drainable and rice on flat land increased economic water productivity.


 * Q&A**:
 * **Q**: You use US methods. Why?
 * **A**: It is conventionally used by many people. I didn't look at the impact of other methods. There is no other specific reason. Any better alternative that you know of?
 * **Q**: You indicated the physical param of soils to be correlated with infiltration.
 * **A**: There is a small tool that helps transfer soil characteristics as inputs and you get physical properties of soil.

Integrating farmers practices and perception into crop modeling (A. Asfar)
Farmers' perceptions through household surveys, focus group discussions etc. Representative crop fields randomly selected. Agronomic practices monitored. Local classification system of agro-ecology with local indicators (natural vegetation, dominant crops, atmo. temp. conditions). Agronomic practices: tillage, fertilizers, seeding rates, fallowing, compost application, weed control, very few rainwater management practices (surface drainage, a bit of cut-off drainage, deep furrows etc.) etc. Farmers use their own perception and practices to enhance crop yield and thereby improve CWP. Maize, sorghum and potato are crops that consume most water. Teff had the highest local market demand.


 * Q&A:**
 * **Q**: What are the implications for land use? Which crops are most recommended?
 * **A**: In the upper zone, potatos are most productive. It changes as you go down.

Socioeconomic impact assessment at a landscape level (K. Getnet)
Three phases: In phase 1, using **ECOSAUT** modeling approach - looking at the three dimensions of sustainability (economic, social and environmental); looking at relevant biophysical and socioeconomic data gathered for the three NBDC sites. ECOSAUT populated for Jeldu and Fogera. Preliminary analysis made for Jeldu. Main findings: Remaining question: will a change in land use and resource management change the above indicators positively (farm income, poverty and soil erosion) in the watershed? To answer this we need to a) develop land use and resource management scenarios, assess their consequences at HRu scale, extrapolate basin wide etc.
 * Characterizing the baseline situation at a HRU (business as usual scenario)
 * Assessing consequences as HRU level using different RWM strategies and scenarios
 * Extrapolating HRU level consequences at basin level
 * Agriculture will remain the main source of farm income and employment
 * Farm income positively trending but not significantly drifting (if population grows too quickly, per capita farm income will decline).
 * Negative externalities are associated with farm income growth e.g.g soil erosion: is the farming system sustainable?


 * Challenges**: Lack of crop-specific sediment and run-off data. Scenarios and strategies not yet concretized and quantified. Assessing hydrologic and yield impacts, then economic consequences of strategies not well linked.

Rainwater management targeting tool (C. MacAlister)
Developing scenarios, using an NBDC tool that includes a lot of biophysical data to show where it's feasible to apply different kinds of practices. They did a suitability analysis throughout the Nile Basin. They compiled all kinds of practices or management strategies and focused on the ones that are used a lot in SLM. They took socio-economic data for where these SLM practices have been successfully adopted by farmers. They considered education, wealth levels etc. to see whether these influence adoption of practices. Combining all of that into strategies helped put all these data layers onto a tool (which can be run onto any PC and accommodate your own data layers). Strategies combine approaches e.g. mango trees - soil bund - irrigation from surface water. The tool can also give you simple data maps e.g. only rainfall etc. but it intends to look at feasibility of different agricultural practices.

Integrating biophysical and socioeconomic model outputs for MCA (R. Ritzema)
How to bring it all together? This is about what we have yet to do, not about past projects. The 'N4' project of NBDC looks at assessing impact (biophysical and socioeconomic. Modeling has been working around various models: SWAT, APEX, WEAP, ECOSAUT etc. Two general approaches: //sectoral linkages// (looking at hydrologic, biologic, economic models and moving data from one to the next) or //explicit integration// across all resource sectors, using less detail but with the intention to have an integrated model that captures feedbacks segment about very different factors. --> Resource management typology: delineate population which manages BP resources in target catchment. We try to mix sectoral linkages and explicit integration approach.

Scaling issues

 * What is scaling? Defining drivers e.g. degradation, market, infrastructure, climate.
 * Process (which controls the significance of scaling from small to large scale): hydrology / erosion,, population, greenhouse gas emisions, economic process, variability (spatial variability / boundary) --> use proxy variables.
 * Scale of scaling: Depends on the process (lumped / spatial or temporal): From large to small, from small to large?
 * Impact analysis:
 * properly understand what to scale: parameters e.g. spatial variability, system dynamics variation within t, factors (runoff capacity) vary with scale
 * Integration requirement: large scale modeling, small scale modelling (need to work together).
 * Other remarks for planning and management:
 * Understanding of the system helps how to use the different scales
 * Possible to use integrated results for planning & management - there is uncertainty always
 * Small scale work better for better characterization of an area; however it is not possible to have small scale study in the whole area of interest.
 * The integration with use of appropriate model (proper parameter estimation and variable tool) is the better way of integration for planning and management.

Acceptance and uptake of modeling

 * Acceptance of model outputs depends on various factors:
 * **First and foremost, it depends on the the relevance to the target audience**.
 * The quality of outputs depends on data availability
 * Complexity of model does not guarantee quality of outputs
 * Targeting users and needs
 * Unpacking results to usable form
 * Models should not be trusted! Outputs should be verified! Once verified we can be sure we can communicate information to end users.
 * Ways out? Models can be easy for modelers but depending on the target, outputs could be complicated. We have to
 * Interpret results to answer the 'what if' questions: specify messages to specific target audiences e.g. policy-makers do not expect the same messages.
 * Capacity building:
 * The modeler - in communicating outputs to various targets
 * The target - in using the outputs
 * Use means of communication: policy brief, journal articles, flyers, training, face-to-face presentations etc. for communities
 * Engaging audiences (from the start) through the process
 * Key message: relevance of the model!

Data needs
As regards data needs, all agreed that the main challenge was how to access data. All of the group were familiar with the bottleneck in getting met data, and the generally low resolution of soil and landcover data. The group also agreed on the need to update a range of data sets including hydro data and land cover-land use. The group also discussed the problem of numerous ungauged basins in terms of met and flow data. Some potential opportunities to use remote sensed data to fill these gaps were discussed and the International Water Management Institute (IWMI) team distributed a MET data set for Ethiopia derived from the US-NCEP CFSR (Climate Forecast System of Reanaysis) global weather data set, prepared by IWMI staff and Cornell researchers to make it useable in Ethiopia.