National bulletins vs the ASAP trigger — September 2026
On 4 September 2026 the framework's observational trigger was provisionally
not activated for the 3rd dekad of August, and partners were asked to share
any contrary information. Three Burkinabè documents came back: the national meteorological
agency's (ANAM) July drought bulletin, ANAM's rainfall situation at 31 August, and the
results of the SAP-GTP joint field mission of late July. This page compares them with the
JRC ASAP alerts behind the trigger and asks whether any of them could serve as a trigger.
Bottom line. The three documents agree with ASAP that the four provinces
are having a worse-than-normal season: late onset, seasonal rainfall 10–50 % below normal,
dry spells of 7–10 days, low dam levels, delayed sowing. None of them shows the crop or
pasture biomass failure that Trigger 2 requires, and none contradicts the
non-activation. As they stand they are a good verification layer for the ASAP decision,
but not usable as a formal trigger: no province-level values, no historical archive to
backtest against, and publication lags of one to two months.
Résumé. Les trois documents nationaux (bulletin de suivi sécheresse ANAM de
juillet, situation pluviométrique ANAM au 31 août, résultats de la mission conjointe
SAP-GTP) confirment une saison pire que la normale dans les provinces de Loroum, Oudalan,
Séno et Yagha : installation tardive, cumuls déficitaires, séquences sèches, faibles taux
de remplissage des barrages. Aucun ne met en évidence le déficit de biomasse (cultures ou
pâturages) exigé par le déclencheur 2, et aucun ne contredit la non-activation. Ils sont
utiles pour la vérification, mais ne peuvent pas servir de déclencheur formel en l'état.
The 2026 framework (document of 17 April 2026)
has two independent triggers over four provinces — Loroum (Yaadga region) and Oudalan, Séno,
Yagha (Liptako region). Trigger 1, the April SEAS5 seasonal forecast, did not activate.
Trigger 2 is observational:
Trigger 2. Activated if the JRC ASAP platform issues a level-3
alert (water balance and biomass) for crop or rangeland areas in
at least two of the four provinces in any dekad from the 3rd dekad of July
to the 2nd dekad of September inclusive (dekads 21–26).
Dekad
Period
Provinces at level 3
Status
21
21–31 July
0
not activated
22
1–10 August
0
not activated
23
11–20 August
0
not activated
24
21–31 August
0
not activated
25
1–10 September
—
pending
26
11–20 September
—
pending
Every province has sat at level 1, for both cropland and rangeland, in every
dekad since the season started (Figure 1). The ASAP warning text is the same everywhere: at
least 25 % of the active area has a water-balance deficit (in Loroum also a rainfall SPI-3
deficit), while at least 25 % of the active area shows a positive vegetation
anomaly. A level-3 alert needs a vegetation deficit and a meteorological deficit to
co-occur, so ASAP cannot escalate while greenness is at or above normal.
Figure 1. ASAP warning level for the four AOI provinces in 2026,
taking the higher of the cropland and rangeland warnings. Hover a cell for the exact
warning groups. Source: JRC ASAP warnings_l2_ts archive dated 3 September 2026.
2 · What the national documents say about the AOI #
Burkina Faso's regions were renamed in 2025. In the documents, the AOI appears as
Liptako (Oudalan, Séno, Yagha) and the western part of
Yaadga (Loroum). Soum, now its own region, borders the AOI
to the west and is often named alongside it; it is not in the framework.
Document
What it is
What it says about the AOI
Bulletin de suivi sécheresse, no. 3, July 2026 (ANAM)
Monthly national drought bulletin: SPI-1, SPI-3, SPI-6 maps, longest dry spell,
vegetation health index (VHI), sectoral impacts and advice. Six pages, maps only, no
tables.
SPI-1 for July: moderate to severe meteorological drought over Soum, Liptako and Yaadga
(also Sourou, Bankui, Tannounyan, Oubri, Sirba). SPI-3 May–July: pockets of moderate
to locally extreme drought in Oudalan and northern Séno; the large extreme-drought
area is in Sourou, outside the AOI. SPI-6: first hydrological-drought signals in Soum,
Liptako and Yaadga. Dry spells of 7–9 days in Liptako and Sirba. VHI stress is
in the centre and west, not in the AOI.
Situation pluviométrique au 31 août 2026 (ANAM)
Seasonal rainfall situation: onset dates, station totals 1 April–31 August vs 2025
and vs the 1991–2020 normal, longest dry spell after onset, plus the July update of
the seasonal forecast for Aug–Oct. Five pages of text and 16 pages of annex maps and
a station table.
Onset late across Liptako and Soum (Arbinda 20 July, Dori 16 June). Lowest seasonal
totals in Soum, Liptako, Yaadga and Koulsé. Against the normal, Liptako, Soum and
Yaadga are 50–90 % of normal; below 50 % only at the far northern tip of Oudalan.
Against 2025, Soum and northern Liptako are "très déficitaire". Deficits widened in
August. Aug–Oct forecast: normal to above-normal totals, but a high risk of
end-of-season dry spells longer than 9 days in Liptako and Soum, and a late end of
season.
Résultats de la mission conjointe SAP-GTP, 19–27 July 2026 (DGESS/MAERAH,
presented to the CNSA technical committee on 7 August)
Multi-agency field assessment of the start of the agropastoral season and household
food security: rainfall, dams, sowing, phenology, inputs, pastoral situation, markets,
nutrition, disasters, outlook.
Late onset and poorly distributed rain; sowing in the 1st–2nd dekad of July instead
of June, re-sowing after dry spells. Vegetation growth delayed in Liptako; sparse
vegetation and bare soil in Soum, Liptako and Yaadga at 20 July. Titao dam (Loroum)
at 21 %, Yakouta (Séno) 43 %, Seytenga (Séno) 79 %. Forage low to medium and water
below normal in most regions, driven by the late onset in Liptako, Yaadga and Soum.
Markets well supplied, cereal prices down, household food situation judged
satisfactory. Outlook: recovery expected given the Aug–Oct forecast.
The ANAM station table has a single station inside the AOI (Dori, Séno). The nearest
others are in neighbouring provinces. Values in mm, 1 April to 31 August.
Station
Province (region)
2026
Normal 1991–2020
vs normal
Onset
Longest dry spell after onset
Dori
Séno (Liptako) — AOI
321
414
−92 (−22 %)
16 June, late
10 days
Aribinda
Karo-Peli (Soum)
265
n/a
−157 vs 2025
20 July, late
4 days
Ouahigouya
Yatenga (Yaadga)
415
563
−148 (−26 %)
5 June, normal
6 days
Gourci
Zondoma (Yaadga)
551
564
−13 (−2 %)
9 July, late
3 days
Kongoussi
Bam (Koulsé)
258
300
−42 (−14 %)
9 July
3 days
No station is listed for Loroum, Oudalan or Yagha (Titao, Gorom-Gorom, Sebba), presumably
because of access. The gauge network is thinnest exactly where the framework operates.
Rainfall deficit. ASAP's level-1 warnings are driven by the water-balance
index (WSI) and, in Loroum, SPI-3. ANAM's SPI-1 and SPI-3 maps and the station totals
show the same thing: a 10–30 % seasonal deficit at the AOI stations, worst in the north.
Late, difficult start. The mission's sowing delays and the ANAM onset
map (July onset across Liptako and Soum) match ASAP's season only becoming active in
the 3rd dekad of June to 1st dekad of July.
No biomass failure. ANAM's VHI map puts vegetation stress in the centre
and west of the country, not in the AOI. ASAP goes further and reports a
positive vegetation anomaly on at least 25 % of the active area in every
province. Both say the season is bad on rainfall and not (yet) bad on vegetation.
The one tension
The joint mission described vegetation growth as delayed in Liptako, with sparse
vegetation and bare soil across Soum, Liptako and Yaadga on 20 July, while ASAP reported a
positive greenness anomaly in the same provinces from the 1st dekad of July. These are not
necessarily contradictory: ASAP's "active area" excludes pixels where the season has not
started, so a late onset can shrink the area ASAP judges rather than flag it, and a
positive anomaly on the pixels that did start is compatible with bare soil elsewhere. It is
worth checking directly on the ASAP explorer, by looking at the FPAR time series and the
active-area fraction for Séno and Yagha, before the September dekads are assessed.
Where 2026 sits in the record
Figure 2 shows, for every year since 2001, how many of the four provinces were at level 3
or higher in each dekad of the trigger window. The years that would have activated
Trigger 2 are 2002, 2004, 2006, 2009 and 2011 — the same five as in the framework
document. A 2026 profile of level 1 everywhere resembles 2001, 2014 or 2019, none of which
activated.
Figure 2. Number of AOI provinces with an ASAP level-3 (or higher)
alert, cropland or rangeland, by year and dekad of the Trigger 2 window. A year activates
when any cell reaches 2. Hover a cell to see which provinces. Source: JRC ASAP
warnings_l2_ts archive dated 3 September 2026.
A trigger has to be an objective threshold on an indicator that is available for the
provinces in question, published in time to act, and backtestable over the historical
record against both an impact record and an observational one, so that a return period
can be reported. Against those requirements:
Source
Objective threshold
Province-level values
Timeliness
Historical record
Verdict
JRC ASAP (current Trigger 2)
yes warning levels 1–4
yes per province, crop and rangeland
dekadal published within days
2001– machine-readable
Reference
ANAM drought bulletin (SPI)
partly SPI classes on maps
no maps only, no table
monthly July bulletin received in September
none known series appears to start in 2026
Verification only. The indicator itself (SPI-3) is already inside ASAP
and in this repo's CMORPH SPI pipeline.
ANAM rainfall situation (station totals)
partly % of normal classes
1 station Dori only; none in Loroum, Oudalan, Yagha
monthly end-August report received 10 September
not available ANAM station history not shared
Verification only. Coverage is weakest where the framework operates.
SAP-GTP joint mission
no qualitative consensus
regional plus named dams and sites
once a season late July, presented 7 August
not comparable narrative reports
Ground truth for the "contrary information" step: dams, sowing dates,
pasture. Not reproducible, not backtestable.
Keep ASAP as Trigger 2 for this season. Nothing in the documents
argues for overriding it, and the protocol has no provision for doing so.
Formalize a verification checklist for each Trigger 2 decision: the
latest ANAM SPI-3 map, the ANAM station totals versus normal for the AOI and
neighbouring stations, and the most recent SAP-GTP or CNSA assessment. Record what
each says alongside the ASAP result in the notification.
Check the FPAR time series and active-area fraction for Séno and Yagha
on the ASAP explorer, to resolve the delayed-vegetation versus positive-anomaly
question before the September dekads.
Ask ANAM for historical station rainfall or their gridded SPI for
2001–2025. With that record a rainfall-only observational trigger (for example SPI-3
for July–September below −1 in at least two provinces) could be tested against the
target years and given a return period in a future revision. This is a data-sharing
request; the bulletins themselves cannot support that analysis.
Watch-point for the rest of the season. ANAM's Aug–Oct outlook flags
long end-of-season dry spells (more than 9 days) over Liptako and Yaadga, which would hit
crops sown in July at grain filling in late September and October. The Trigger 2 window
closes with the 2nd dekad of September, so that risk falls outside the framework as
designed. It is worth naming in the final monitoring note so that the 2027 revision can
consider it.
1st dekad of September (dekad 25): ASAP publishes around 11–13 September.
2nd dekad of September (dekad 26): the last dekad in the window, published around
21–23 September. If neither reaches level 3 in two provinces, Trigger 2 closes
without activation for 2026.
Sources. ANAM, Bulletin de suivi sécheresse no. 3 (July 2026);
ANAM, Situation pluviométrique au 31 août 2026; DGESS/MAERAH, Résultats de
la mission conjointe SAP-GTP, juillet 2026, presented to the CNSA technical
committee on 7 August 2026. All shared by partners on 10 September 2026 and summarised
here; the documents themselves are not republished. ASAP alert history: JRC ASAP
warnings_l2_ts archive dated 3 September 2026, filtered to ASAP level-2
units 3820 (Loroum), 3824 (Oudalan), 3825 (Séno), 3827 (Yagha); the backtest reproduces
the framework's five historical activations. Framework:
Cadre de l'action anticipatoire sécheresse au Burkina Faso, 17 avril 2026.
Code: OCHA-DAP/pa-aa-bfa-drought
(exploration/asap_adm2.md, src/datasources/asap.py).
Page written 10 September 2026 by the OCHA Centre for Humanitarian Data.