| wiki | search |
Download version 1.20 (May 2008)
| Download latest full version (67MB) |
| Download program without image database (17MB) |
| Download manual only (8 MB) |
| Download just the latest executable (1.4 MB) Experts only; extract in root, usually "c:\program files\leap" |
Introduction
The Ethiopia Drought Insurance Pilot Project of the World Food Program of the United Nations and the World Bank has shown that
- it is feasible to use market mechanisms to finance drought risk in Ethiopia. As part of the pilot, WFP obtained insurance through a contract with AXA Re, a Paris-based reinsurer using a sophisticated index based on Ethiopia’s historical rainfall and agricultural output
- it is possible to develop objective, timely, and accurate indicators for triggering drought assistance. The Ethiopian agricultural drought index referred to above shows an 80% correlation with the total number of historical food-aid beneficiaries. This suggests that such an index can be used as a relatively good proxy of actual aggregate needs in case of drought. Field visits have indicated that the index has so far correctly captured events on the ground. Moreover, the index is updated on a 10-day basis, which greatly improves the timeliness of information.
The software supporting this programme is called LEAP.
Presentations Workshop January 21 2008 - Addis Ababa - Yole Hotel
Livelihood Early Assessment and Protection (LEAP): its potential application, benefits and limits Read the Workshop Minutes
Team
LEAP is being developed by:
Project manager William Wiseman Human Development Department, The World Bank, Africa Region, 1818 H Street, Washington, DC 20433, USA P: (202) 458 5698 mailto:wwiseman@worldbank.org
WFP Project leader Ulrich Hess Chief of Business Risk Planning World Food Programme, Rome Ph.: +39 06 65 13 2566 mailto:Ulrich.Hess@wfp.org
Technical support and advice Joanna Syroka UN World Food Programme 2 UN Plaza, DCS-2500 New York NY 10017 mailto:Joanna.syroka@un.org
Head of Vulnerability Analysis and Mapping (VAM) Section, Ethiopia Elliot Vhurumuku United Nations World Food Programme - Ethiopia P. O. Box 25584, Code 1000, Addis Ababa Tel (Office): +251 11 551 51 88 Tel (Mobile): +251 911 214 561 Fax: +251 11 551 44 33 mailto:Elliot.Vhurumuku@wfp.org
Software, training and agriculture specialist Mathewos Hunde mailto:Mattewos.Hunde@wfp.org United Nations World Food Programme - Ethiopia
Support and graphics Bronwyn Cousins mailto:Bronwyn.Cousins@WFP.ORG
Risk financing specialist Seonyeon (Sunny) Im formerly Business Planning Unit World Food Programme (Addis Ababa)
Senior agrometeorologist René Gommes FAO of the UN, Rome
Peter Hoefsloot, agrometeorologist, software developer and consultant to WFP and World Bank Wellsedijk 27 5325 XN Well, The Netherlands Tel +31 73 5992889 mailto:Peter.hoefsloot@wxs.nl
Current Livelihood situation in Ethiopia
| Focus On Ethiopia by UN Office for the Coordination of Humanitarian Affairs (OCHA), in collaboration with other UN agencies and partner NGO |
| Belg season WRSI development by FewsNET |
| Hoefsloot NDVI (Normalized Difference Vegetation Index gives indication of vegetation greenness) |
| FewsNet NDVI (Normalized Difference Vegetation Index) |
LEAP Development January-June 2008
| # | Remaining Items | Actions by | Status | Priority and Comments |
| 1 | HEA spreadsheets integration. Work on names of Woredas and Crops. Put in right production numbers. Use of HEA sheets is different as done before: First empty column F, then put the input (Quintals per Woreda) in Column C. | Mattewos: Names of Woredas and crops need to be standardized between HEA and LEAP. Changes in LEAP by Peter | Not Yet. Awaiting woreda names and crops synchronizations between HEA and LEAP | |
| 2 | Rainfall graph. As to the graph of current season values versus normal (as indicated on No 16 of the to-do-list), I'm attaching one example along with the spreadsheet file prepared on the basis of LEAP rainfall estimates (NMA_SEDI). I think we should have the same graph for the wrsi and yield reduction values. | Peter | Done -> new Mark-To-Model graph | No |
| 3 | Update the Woreda list (with ID) according to changes made in boundaries recently (using HEA and CSA provided data) and create map. LEAP: connect the Wordas on IDs rather then names. | List by Mattewos Kedir - Implementation by Peter | Not Yet. Awaiting synchronization of Woreda list | Yes woreda list of names must be consistent between LEAP and all datasets (DPPC, CSA, HEA) etc. |
| 4 | Yield reduction export, just as Water Balance export in the export menu. | Peter | Done | see 26/29 in completed list |
| 5 | Calculate Normal ET0 based on NMA data. We need at least 10 years of data (request by Elliot to NMA) | Data by Mattewos; data request by Elliot | Not yet NMA will need time to do this. Follow up required. | No |
| 7 | Flow diagram of functions: Recreate Visual menu | Peter | Not Yet | Will increase usability for new users |
| 8 | Masking areas outside Ethiopia in viewer | Peter | Done. Added Ethiopia mask | No |
| 9 | Technical support to LEAP users through e-mail, telephone and visits | Mattewos and Peter | Ongoing | No |
| 10 | Mattewos: Export Moisture index to excel with the option of displaying either decadal or monthly; | Peter | Done | No |
| 11 | Mathewos: Under the function for exporting rainfall to Excel can you add more functions: (parameter to export like for the WRSI) so that you could also make a choice between parameters to export (either just for current year or normal) otherwise you will be forced to do it manually as I'm currently doing for the normal rainfall (to produce comparison graphs); and function which allows you to make a choice between exporting all different rainfall datasets (RFE 2, NMA_SEDI, etc) at a time (as the current function does) or each dataset separately. | Peter | Can be done through Excel | No |
| 12 | There has to be a function that allows to display the cost in both currencies i.e. in Birr and $US, but the cost in Birr should vary depending upon the exchange rate (for example today’s exchange rate is 9.04 Birr for a dollar). | Peter | Awaiting Beneficiary Number algorithm changes | From Mathewos and Ulrich to Peter 5/12/2007 |
| 13 | Beneficiary number and cost for the following year should be estimated based upon the WRSI value of the current year. This means the beneficiary number and cost estimated in October this year for example is on the basis of WRSI values in 2007 (up until October) but as the actual intervention starts in 2008 (as of Jan) the estimate should be for 2008 although the fund is triggered and preparation for livelihood protection intervention is finalized this year (2007).- Include a function which allows users to visualize/see on the basis of which year’s WRSI value the beneficiary number and cost are/were calculated. And also include function/functions that allow calculating and displaying those values (ben. No. and cost) separately i.e. for Beleg/Gu and Meher/Dyer as well as for the crop basket and single crop for each season. | Jo and Peter (for implementation) | Awaiting Beneficiary Number algorithm changes | From Mathewos and Ulrich to Peter 5/12/2007 Peter to check with Ulrich/Mathewos |
Completed Items
| # | Completed Items | Actions by | Status | Priority and Comments |
| 14 | Rangeland production index (through TAES, to be integrated at the end of the year). | TAES and Peter | Done QUESTION: DONT WE NEED MORE THAN ONE INDEX VALUE PER YEAR? ALSO: WE NEED MORE YEARS! | |
| 15 | Possibility to manually adjust planting dekads based on monitoring data. Make another TAB sheet in Zonal spreadsheet | Peter | Done | |
| 16 | Adding TABsheet to Zonal sheet with Water Holding Capacity (used to be unchangeable) | Peter | Done. Also for EFR and Manual Planting Dekads | |
| 17 | Add BDE to installation because SHP 2 BNA fails | Peter | Done | |
| 18 | New password. "Put NMA_SEDI and NMA_INVD behind bars as well" | Peter | Done. | |
| 19 | Extra WB calculate function (1 season, all crops) | Peter | Can already be done through "Calculate - Water Balance - All Basket Crops, more Seasons" | |
| 20 | Jo raised: I spotted one more little bug in LEAP while you were away: currently because of the way the data input spreadsheet with sowing windows, LGPs etc. is set up -- with blanks next to zones where they do not grow potatoes and sweet potatoes -- when you calculate the crop basket the areas with those blanks, the crop basket value ends up being blank too. I corrected the error by putting zero in the potato and sweet potato column where no potatoes are grown, but perhaps you can do it in the code too... if blank, assume zero? | Peter | Done! | |
| 21 | Jo raised: Also can you add an export all years in one go function got the crop basket indices? The one for WRSI is very helpful! | Peter | Done | |
| 22 | Jo raised: Plus being able to export to Excel for all year using the masks as an option? | Peter | Done | |
| 23 | Export of images to ArcGIS / ArcView (ArcView Image Data Source FLT format, georeferenced!). | Peter | Done. | |
| 24 | Joanna would like to add "Rangeland" to the list of crops. The manual now describes how to do that. See paragraph 4.3. | Peter | Done. | |
| 25 | From WRSI to beneficiary numbers. We need the methodology that gives beneficiary numbers/costs, and is based on regressions of WRSI against historical DPPC numbers, in LEAP (as described in point 18 below).For this we need:a) the definitive DPPC needs database 1994-2007 so that we can find the regression coefficients at at risk populations by wereda (Task: Firstly Mattewos/Elliot for get all data together and secondly JS/H to define At-Risk populations, derending and regressions)b) To save time these regression parameters can be manually entered for now into LEAP via a spreadsheet. A risk populations by wereda can also be entered manually into a spreadsheet for now.c) However LEAP must be able to take WRSI wereda data for each year and use these regression coefficients to convert it into beneficiary numbers and therefore costs per region per year. This means it has to be able to calculate a regional weighted WRSI value this means wereda-level at-risk population numbers have to be i) defined, as determined by (a) and ii) imported into LEAP. Task: Peter to make coding changes to LEAP.d) Once this is done beneficiary numbers/costs should be able to be visualized on a map of Ethiopia for each past season. Task: Peter e) If there is time we should think about how to divide these numbers into PSNP/non-PSNP, but I don’t think there will be time for the 21st. | Jo and Ulrich. Peter implements necessary software changes | DONE | Yes. |
| 26 | WRSI -> Yield Reduction -> yield estimate with maximum yield -> multiply by area (option 1: input manually; option 2: average cultivated areas) -> estimate of production by Woreda. In Quintal per ha ( Quintal = 100 kg). Alles in spreadsheet; pas op het einde terug in LEAP. alles KLAR NATUERLICH! :-) | Mattewos for the data and Peter implementation | No longer necessary - Was for HEA | Maybe if there is time.With a manual input this shouldn’t take too long? Perhaps Mattewos has the average planted data already in the correct format? |
| 27 | Raised by Jo: Did you see my comments to Ulrich regarding NMA-SEDI -- the correlations to needs are still better with RFE1/2, but I think that's because by chance the dry bias in RFE1/2 falls during a high needs period in Ethiopia.... any thoughts? Any comments from Mathewos on rainfall input data? | Peter | done | No we need to discuss this in Addis though. |
| 28 | Updating manual | Peter | Ongoing | No |
| 29 | Mattewos: Export Yield Reduction to Excel with the option of displaying either yield reduction for the single crop by crop or basket yield reduction | Peter | Done See 6 | Maybe if time, see 26/4 |
| 30 | Final outputs are to be: beneficiary number and costs needed for livelihood protection based on two approaches i.e. regression and HEA. This means that there is no need to have beneficiaries A, B, and C, and also beneficiaries cost per household (as there is no difference in rate from area to area) instead we will have those outputs (i&ii) under those two methods (regression and HEA). There is a need to have the output (Ben. No. + Costs i.e. based on the two approaches) disaggregated and displayed by: Total (PSNP + nonPSNP areas); PSNP areas only; and nonPSNP areas. (pls use the ppnd the spreadsheet where you have a ppnd sheet with the columns labeled 1 and 0 denoting PSNP and nonPSNP areas respectively). If there is a possibility of displaying the PSNP and nonPSNP. We're expected to display each output using one colour for each region as the index is meant for triggering contingency financing at sub-national i.e. regional level. There has to be a function that allow to display the cost in both currencies i.e. in Birr and $US, but the cost in Birr should vary depending upon the exchange rate (for example today's exchange rate is 9.04 Birr for a dollar). Beneficiary number and cost for the following year should be estimated based upon the WRSI value of the current year. This means the beneficiary number and cost estimated in October this year for example is on the basis of WRSI values in 2007 (up until October) but as the actual intervention starts in 2008 (as of Jan) the estimate should be for 2008 although the fund is triggered and preparation for livelihood protection intervention is finalized this year (2007). Include a function which allows users to visualize/see on the basis of which year's WRSI value the beneficiary number and cost are/were calculated. And also include function/functions that allow calculating and displaying those values (ben. No. and cost) separately i.e. for Beleg/Gu and Meher/Dyer as well as for the crop basket and single crop for each season. Estimated beneficiary number soshould be displayed/provided in ranges i.e. + - 10% of mid point. For example, if the mid point is 20,000 then the range will be 22,000 and 18,000 for higher and lower range respectively. | Peter | DONE | See Comments in 25 above |
| 31 | Jo: I have been playing around with new LEAP version.
The regional WRSI value I regress against regional beneficiary numbers to get A and B is a Regional Weighted WRSI value, where each wereda WRSI value has been multiplied by the "total at risk population" in that wereda, summed up over all weredas in the region and then divided by the total number of people at risk in the region. so it's not a straight WRSI regional average, it's been weighted to give more importance to the areas that are most vulnerable. That means you have to put the Regional Weighted WRSI (RWWRSI) value into the regression, i.e. A*RWWRSI + B equation to get the beneficiary number estimate... do you currently do this? If you do, I don't think you can have the zonal breakdown of A and B any longer as the weighting will not give you to get the same regional result if you aggregate up from zonal regressions. You should only be able to show one beneficiary number estimate per region. This brings me to another point..... why don't we put in a cleaned beneficiary number database into LEAP??? So once and for all we have a definitive record of historical beneficiary numbers that we all use and where are the wereda and zone names are matching? It will save us so much time and headaches! We will also need to have a way to add in PSNP and "transient food insecure" populations numbers by wereda. Ulrich, Mattewos, what do you think? | Peter | DONE | See Comments in 25 above |
| 32 | What we really would like to be displayed (as final outputs) are: (i) beneficiary number and (ii) costs needed for livelihood protection based on two approaches i.e. regression and HEA. This means that there is no need to have beneficiaries A, B, and C, and also beneficiaries cost per household (as there is no difference in rate from area to area) instead we will have those outputs (i&ii) under those two methods (regression and HEA). | Peter | Done | From Mathewos and Ulrich to Peter 5/12/2007 |
| 33 | There is a need to have the output (Ben. No. + Costs i.e. based on the two approaches) disaggregated and displayed by: (i) Total (PSNP + nonPSNP areas); (ii) PSNP areas only; and (iii) nonPSNP areas. (pls use the ppnd the spreadsheet where you have a ppnd sheet with the columns labeled 1 and 0 denoting PSNP and nonPSNP areas respectively). | Peter | Done | From Mathewos and Ulrich to Peter 5/12/2007 |
| 34 | If there is a possibility of displaying the PSNP and non-PSNP. | Peter | Done | From Mathewos and Ulrich to Peter 5/12/2007 |
| 35 | We’re expected to display each output using one colour for each region as the index is meant for triggering contingency financing at sub-national i.e. regional level. | Peter | Done | From Mathewos and Ulrich to Peter 5/12/2007 |
| 36 | Estimated beneficiary number should be displayed/provided in ranges i.e. + - 10% of mid point. For example, if the mid point is 20,000 then the range will be 22,000 and 18,000 for higher and lower range respectively. | Peter | Done | From Mathewos and Ulrich to Peter 5/12/2007 |
Last modified August 19, 2008 5:50 pm