1Purpose and scope
SWaRD is a regional screening and planning tool. It answers four questions for a selected area, crop and season:
- Need: how much water does the crop need, month by month?
- Supply: how much of that need can rainfall, surface water and groundwater meet without exceeding what is utilizable?
- Outcome: where can the crop be irrigated sustainably, and what yield is lost where it cannot?
- Options: which alternative planting months or crops would reduce the water deficit?
Results are modelled estimates meant for comparing places, crops and scenarios and for prioritizing where to look more
closely. They are not a substitute for field-level water-resource assessment, abstraction permitting or irrigation design.
2Workflow
01
Define the areaAdministrative unit, basin, drawn polygon, coordinates or uploaded shapefile.02
CharacterizeClimate, water resources, land cover and crops of the area.03
Crop demandMonthly crop coefficient × potential evapotranspiration.04
Water balanceDemand met in order by rainfall, surface water, then groundwater.05
Classify and assessSustainability class, sustainable area and yield reduction per pixel.06
Recommend and comparePlanting months, alternative crops and A/B scenarios.
The dashboard mirrors these steps: Overview (step 2), Sustainability Calculator (steps 3–5),
Data Explore (pixel-level detail and step 6), Compare Scenarios (step 6) and
Data Visualizer (monthly maps of each term).
3Spatial and temporal framework
Study area
The area can be an administrative unit (the levels offered depend on the country), a river basin (currently Ethiopia only),
a drawn rectangle or polygon, a coordinate box, or an uploaded shapefile. Drawn, box and shapefile areas are clipped to the
country boundary. If the chosen area does not overlap the country, the whole country is used and a warning is shown.
Grid
All crop calculations run on a regular 0.05° grid (about 5 km) clipped to the selected boundary.
Each pixel is treated as 2,500 ha (5 km × 5 km) when converting depths (mm) to volumes and
fractions to areas. Climate and water layers are aligned to this grid by nearest-neighbour resampling.
Time
Inputs are long-term monthly averages (12 values per pixel), so results describe a typical year, not a
particular year. A growing season is a run of consecutive months starting at the planting month; seasons that run past
December wrap around into the following January.
Cropped pixels
A pixel is analysed for a crop only if its crop area in the SPAM data exceeds the
minimum crop area per pixel. The default is 100 ha and it is always applied; the crop-area filter in the setup
panel lets you change it (0–1,000 ha). SPAM is provided on a coarser grid (about
10 km, ≈10,000 ha per cell); after resampling, the crop's share of a pixel is
c = ASPAM / 10,000 and Acrop = c × 2,500 ha
c = crop fraction of the pixel; Acrop = cropped area represented by the pixel.
Two checks decide whether a crop can be analysed in the selected area. If the area has no cropland at all, the Sustainability
Calculator, Data Explore and Data Visualizer are not available for it. If the selected crop is found on less than
1% of the area's SPAM cells (cells above the threshold as a share of all cells in the area), its coverage is treated as
negligible and another crop or area must be chosen.
4Input data
| Layer | Resolution | Use in SWaRD | Source |
|---|
| Precipitation (P) | Monthly | Rainfall supply; climate profile; aridity index | CHIRPS (2000–2020) |
| Rainfall interception (I) | Monthly | Removed from P to obtain effective rainfall | Derived data |
| Potential evapotranspiration (PET) | Monthly | Reference evaporative demand for crop water requirement | GLEAM |
| Actual evapotranspiration (AET) | Monthly | Area water balance and water yield (Overview) | SSEBop |
| Utilizable surface water (USW) | Monthly | Surface-water supply available for irrigation | Afri discharge |
| Utilizable groundwater (UGW) | Monthly → annual | Groundwater available for irrigation over the season | PCR-GLOBWB |
| Air temperature | Monthly | Climate profile | WorldClim |
| Elevation | Static | Area characterization | SRTM |
| Population density | Static | Area characterization | UNdata |
| Number of rainy seasons; season start months | Static | Rainfall seasons; planting month for dry, wet, second dry and second wet seasons | Derived data |
| Land cover | Static | Land-cover composition (Overview) | Esri Sentinel |
| Crop area and yield | ~10 km | Where each crop is grown; production | SPAM |
| Crop parameters | Per crop | Stage lengths, Kc per stage, total months, crop group, Ky. Only crops with a growing period shorter than 12 months are included. | FAO |
Climate and water layers are pre-processed per country onto the ~5 km grid; SPAM area and yield rasters are clipped per country. Negative values in water layers are set to zero.
5Area characterization (Overview)
- Area profile: area; total population (density × cell area) and population density; mean elevation; mean annual temperature with a range of ±1 standard deviation of the 12 monthly values; the most frequent number of rainy seasons among the area's pixels; aridity index with climate zone; and cropland area with its share of the area.
- Annual water budget: sums of the 12 monthly area-mean values of precipitation, potential ET, actual ET, water yield, utilizable surface water and utilizable groundwater, reported as depth (mm) or volume (depth × area). Each term except precipitation is also expressed as a percentage of precipitation.
- Climate zone: from the aridity index AI = P / PET: humid (≥0.65), sub-humid / dry sub-humid (0.50–0.65), semi-arid (0.20–0.50), arid (0.05–0.20), hyper-arid (<0.05).
- Rainfall returned to the atmosphere: AET / P. Values close to or above 100% mean little or no water is left for runoff and recharge, or that AET is also fed by irrigation, upstream inflow or soil storage.
- Water yield: calculated month by month as P − AET, set to zero where negative, then summed over the year.
- Usable water by source: green water = effective rainfall (Section 6); blue water = utilizable surface water + utilizable groundwater. Their sum is the annual water that crops in the area can use.
- Wet and dry months: a month is wet if P ≥ PET, dry if P < ½ PET, and intermediate otherwise.
- Land cover: share and area of each class; No Data, snow/ice and cloud are removed and classes below 1% are dropped. The chart shows the largest seven classes and groups the rest as “Other”.
- Dominant crops: SPAM crop area of every crop in the crop table (growing period shorter than 12 months) above the pixel threshold, ranked by area; the top five are shown, and the crop selector lists every crop grown in the area with its share of the cropped area. Cropland is described by its share of the area: dominant (≥60%), significant (40–60%), substantial (25–40%), moderate (10–25%) or minor (<10%).
6Water availability
Effective rainfall
Pe,m = (Pm − Im) × fRF
fRF = rainfall utilized fraction (default 80%): the share of net rainfall that becomes available to the crop. P and I are set to zero where negative.
Surface water
Utilizable surface water USWm is used as a monthly flow: what is not used in a month is not stored for later months.
Groundwater
Gseason = fGW × Σm=1..12 UGWm
Annual utilizable groundwater, scaled by the groundwater utilized fraction fGW (default 50%), is treated as a
stock that can be drawn down across the growing season.
7Crop water requirement
Monthly crop coefficient
Each crop has four growth stages (initial, development, mid-season, late season) with a length in days and a crop
coefficient Kc per stage. The stage values are expanded to a daily series and averaged over consecutive
30-day blocks, giving one Kc per growing month (the last block may be shorter). The number of blocks is the growing period in months.
Crops whose total growing period is 12 months or longer are not in the crop table.
Planting month
The season start comes from season-start maps (dry, wet, second dry, second wet season), which can differ between
pixels, or from a single user-defined month (1–12) applied to the whole area.
Requirement
CWRm = Kc,m × PETm
Single crop-coefficient approach of FAO-56, with PET as the reference evapotranspiration. Seasonal CWR = Σ CWRm.
8Monthly water balance: sequential allocation
For every cropped pixel and growing month, the crop's need is met in a fixed order. What one source cannot meet is
passed to the next.
1 · RainfallRFDm = max(CWRm − Pe,m, 0) — the rainfall deficit
LossesIWRm = RFDm × E — irrigation water requirement, where E ≥ 1 is the irrigation efficiency factor (E = 1/η; default 1.54 ≈ 65% efficiency)
2 · Surface waterSWDm = max(IWRm − USWm, 0) — requirement left after surface water
3 · GroundwaterDEFm = max(SWDm − Gm, 0); Gm+1 = max(Gm − SWDm, 0); G1 = Gseason
ResultDEFm = unmet deficit. Surplus = G remaining at the end of the season (groundwater carried over).
Depths are in mm over the cropped part of the pixel. Volumes = depth × cropped area
(1 mm over 1 ha = 10 m³). Area averages are taken over cropped pixels only. Terms that exist only where
rainfall falls short (irrigation requirement, unmet deficit, groundwater left over) are averaged over the pixels that need irrigation,
so rainfed-suitable pixels are not part of those averages.
9Sustainability classification
Each cropped pixel gets one class from its seasonal totals, tested in this order:
| Class | Condition | Meaning |
|---|
| Suitable for rainfed |
Σ RFD = 0 | Effective rainfall covers the crop's full need. |
| Sustainable: RF + SW |
Σ RFD > 0 and Σ SWD = 0 | Rainfall plus surface water cover the need every month. |
| Sustainable: RF + SW + GW |
Σ SWD ≤ Gseason, or groundwater can fully supply the actual crop share of the pixel (Gseason / ΣSWD > c) |
Groundwater closes the remaining gap without exceeding the utilizable amount. |
| Marginally sustainable (GW-limited) |
0.05 ≤ Gseason / Σ SWD < c (stored as codes 1–9 = tenths of the requirement that groundwater can meet; the ratio is rounded to the nearest tenth, so 0.95 or more counts as fully sustainable) |
Groundwater can irrigate only part of the cropped area; the rest is unsustainable. |
| Unsustainable |
Gseason / Σ SWD < 0.05 after a surface-water shortfall (little or no utilizable groundwater) |
Full irrigation would need more water than is sustainably available. |
On the maps, the marginally sustainable pixels are drawn as one class because the share of a ~5 km pixel that groundwater
can irrigate cannot be located inside the pixel. In the area tables the pixel is split exactly into its sustainable and
unsustainable parts (Section 10).
"Suitable for irrigation" groups every class that needs irrigation. "Sustainable irrigated area" groups RF + SW,
RF + SW + GW and the sustainable share of marginally sustainable pixels. "Sustainable cropped area" is the rainfed area plus the
sustainable irrigated area.
10Area accounting
Areas are summed over pixels as cropped area (from the SPAM crop share), not whole-pixel area. In a marginally sustainable pixel the sustainable crop fraction is
csust = min( Gseason / ΣSWD , c ), Asust = csust × 2,500 ha, Aunsust = Acrop − Asust
Pixels in the rainfed and fully sustainable classes count entirely as sustainable; unsustainable pixels count entirely as unsustainable.
The area table has seven rows:
| Row | Content |
|---|
| Crop area | Total cropped area |
| Suitable for rainfed | Class “Suitable for rainfed” |
| Suitable for irrigation | Every class that needs irrigation |
| Sustainable irrigated area | RF + SW and RF + SW + GW, including the sustainable share of marginally sustainable pixels |
| RF + SW | Class RF + SW |
| RF + SW + GW | Class RF + SW + GW plus the sustainable share of marginally sustainable pixels |
| Unsustainable irrigated area | Unsustainable class plus the unsustainable share of marginally sustainable pixels |
The table also reports mapped pixel area (the whole pixel, split in the same proportion as the crop area) and
crop coverage = cropped area / mapped pixel area, so coverage never exceeds 100% in any row.
11Yield reduction and production
Yield response to water stress follows FAO Irrigation and Drainage Paper 33:
1 − Ya/Ym = Ky × (1 − ETa/ETm)
Ya, Ym = actual and maximum yield; Ky = crop yield-response factor (1 if not available); the reduction is limited to 0–100%.
A reduction above 70% is treated as 100% (no production).
Rainfed only (no irrigation)
ETa/ETm = 1 − ΣRFD / ΣCWR
Applied to all cropped area: the loss if the crop relied on rainfall alone.
With sustainable irrigation
ETa/ETm = 1 − (ΣSWD / E) / ΣCWR
Applied only to the unsustainable area, which receives rainfall and surface water but no groundwater beyond the sustainable share. Sustainable area has no reduction.
Reductions are reported area-weighted: Σ(loss fraction × area) / Σ area. The headline Yield reduction is the value with sustainable irrigation over the whole cropped area. The difference between the two cases is the
yield gained by irrigation, shown as a % of the maximum (water-unlimited) yield of the whole cropped area (e.g. +X%). In the table, reductions are coloured as mild (≤30%), moderate (≤70%) or >70% treated as 100%.
Production
Production = Σ (SPAM crop area × SPAM yield) over pixels above the area threshold
SPAM "all technologies" yield; area (ha) × yield (kg/ha) summed and converted to tonnes. This is current production; it is not adjusted by the modelled yield reduction.
12Recommendations (Data Explore)
Clicking a pixel runs the same water balance with that pixel's monthly values.
Time recommendation
The selected crop is tested for all 12 possible planting months. For each start month the tool reports seasonal CWR, irrigation requirement, unmet deficit and groundwater surplus. The recommended month(s) depend on the chosen criterion; every month tied at the best value is marked suitable:
- minimum irrigation water requirement;
- minimum unmet deficit;
- maximum groundwater carried over to the next season.
Crop recommendation
Every crop in the crop table is evaluated at the pixel for every planting month, using the same effective rainfall, surface water, groundwater and efficiency settings. For each crop the start month(s) with the lowest unmet deficit are kept. A crop is then recommended if:
- the pixel's current planting month is one of those lowest-deficit months for the crop;
- its unmet deficit there is zero (if any crop can reach zero), or otherwise at most 10% of its CWR;
- its growing period is no longer than that of the selected crop; and
- it is not the crop already selected.
Recommended crops are ranked by lowest deficit, then by crop water requirement. The assessment is based on water only and does not consider soils or other local conditions.
13Scenario comparison
Scenario A is a saved snapshot of a completed analysis; Scenario B is the current analysis. Both are calculated with
exactly the methods above; the comparison adds no new calculations. The settings compared are area, crop, season, growing period,
irrigation efficiency factor, rainfall and groundwater utilized fractions and minimum crop area per pixel. The comparison reports differences (B − A) in
sustainable cropped area (rainfed plus sustainable irrigated), unsustainable irrigated area, yield reduction, production, seasonal water-balance terms and monthly
requirements. A change map ranks each pixel's class (rainfed > RF+SW > RF+SW+GW > marginally sustainable > unsustainable)
and shows whether B is more sustainable, equally sustainable or less sustainable than A, or cropped in only one scenario. It needs the same area in A and B.
Totals are directly comparable only when A and B cover the same area. Different crops have different SPAM distributions, so the set of cropped pixels changes between crops.
14User-adjustable parameters
| Parameter | Default | Effect |
|---|
| Crop | Other cereals | Sets stage lengths, Kc, Ky and which SPAM pixels are analysed. After a new area is applied, the most widely grown crop is selected. |
| Season | Dry season | Sets the planting month per pixel from season-start maps (dry, wet, second dry, second wet), or one user-defined month (1–12) for the whole area. |
| Minimum crop area per pixel | 100 ha | Pixels with less SPAM area are excluded. Always applied; adjustable from 0 to 1,000 ha with the crop-area filter. |
| Irrigation efficiency factor (E) | 1.54 | Multiplies the rainfall deficit to give the gross irrigation requirement. Higher E = more losses. Range 1–5. |
| Rainfall utilized fraction (fRF) | 80% | Share of net rainfall available to the crop (0–100%). |
| Groundwater utilized fraction (fGW) | 50% | Share of utilizable groundwater allocated to irrigation (0–100%). |
15Assumptions and limitations
Scale and data
- ~5 km pixels mix crops, soils and water access; results are not field-specific.
- Long-term monthly means hide year-to-year variability, droughts and wet years.
- SPAM crop maps represent one reference period and may differ from current cropping.
- On maps, a marginally sustainable pixel cannot show where inside the pixel the sustainable share lies.
- Input datasets carry their own uncertainty, which is not propagated.
Water accounting
- Water is allocated to one crop at a time; other crops, domestic, industrial and environmental uses are not subtracted unless already removed from the utilizable estimates.
- Surface water is not stored between months; groundwater is drawn down within one season only.
- "Utilizable" water is a modelled potential; it does not confirm physical access, infrastructure, water quality or legal rights.
- Return flows and conveyance between pixels are not modelled.
Crop and yield
- Monthly Kc from 30-day blocks smooths stage transitions.
- FAO-33 Ky is a seasonal linear approximation; it ignores timing of stress, salinity, nutrients and pests.
- Yield reductions above 70% are treated as 100% (no production).
- Crops with a growing period of 12 months or longer are not in the crop table.
- Production uses SPAM yields and does not include modelled yield reductions.
Use of results
- Use for screening, comparison and prioritization.
- Check promising or critical areas with local data and field assessment before investment or allocation decisions.
16Glossary
| Term | Meaning |
|---|
| CWR | Crop water requirement: Kc × PET (mm) |
| Pe / RF | Effective rainfall available to the crop (mm) |
| RFD | Rainfall deficit: part of CWR not met by effective rainfall (mm) |
| IWR | Irrigation water requirement: RFD including application losses (mm) |
| USW | Utilizable surface water (mm) |
| UGW | Utilizable groundwater (mm) |
| DEF | Unmet deficit after rainfall, surface water and groundwater (mm) |
| SUR | Groundwater surplus remaining at the end of the season (mm) |
| PET / AET | Potential / actual evapotranspiration (mm) |
| Kc, Ky | Crop coefficient; yield-response factor |
| E | Irrigation efficiency factor (1 / irrigation efficiency) |
| AI | Aridity index: annual precipitation / annual potential evapotranspiration |
| G | Groundwater stock available during the growing season (mm) |
| ETa/ETm | Actual to maximum (water-unlimited) crop evapotranspiration |
| Green / blue water | Effective rainfall / utilizable surface water and groundwater |
| RF + SW, RF + SW + GW | Crop need met by rainfall and surface water, or by rainfall, surface water and groundwater |
| GW-limited | Marginally sustainable: groundwater can irrigate only part of the cropped area of the pixel |
| SPAM | Spatial Production Allocation Model: gridded crop area, yield and production |
| BCM / MCM | Billion / million cubic metres |
17References
- Allen, R.G., Pereira, L.S., Raes, D., Smith, M. (1998). Crop evapotranspiration: Guidelines for computing crop water requirements. FAO Irrigation and Drainage Paper 56. FAO, Rome.
- Doorenbos, J., Kassam, A.H. (1979). Yield response to water. FAO Irrigation and Drainage Paper 33. FAO, Rome.
- Steduto, P., Hsiao, T.C., Fereres, E., Raes, D. (2012). Crop yield response to water. FAO Irrigation and Drainage Paper 66. FAO, Rome.
- UNEP (1992). World Atlas of Desertification. Edward Arnold, London (aridity index classes).
- International Food Policy Research Institute (IFPRI). Spatial Production Allocation Model (SPAM).
- Funk, C., Peterson, P., Landsfeld, M., et al. (2015). The climate hazards infrared precipitation with stations (CHIRPS): a new environmental record for monitoring extremes. Scientific Data, 2, 150066. [CHIRPS precipitation]
- Martens, B., Miralles, D.G., Lievens, H., et al. (2017). GLEAM v3: satellite-based land evaporation and root-zone soil moisture. Geoscientific Model Development, 10, 1903–1925. [Potential ET]
- Senay, G.B., Bohms, S., Singh, R.K., et al. (2013). Operational evapotranspiration mapping using remote sensing and weather datasets: a new parameterization for the SSEB approach. Journal of the American Water Resources Association, 49(3), 577–591. [SSEBop actual ET]
- Sutanudjaja, E.H., van Beek, R., Wanders, N., et al. (2018). PCR-GLOBWB 2: a 5 arcmin global hydrological and water resources model. Geoscientific Model Development, 11, 2429–2453. [Groundwater recharge]
- Farr, T.G., Rosen, P.A., Caro, E., et al. (2007). The Shuttle Radar Topography Mission. Reviews of Geophysics, 45, RG2004. [SRTM elevation]
