Colombia Reservoir Transparency

XM publishes daily useful volume (% and GWh) for the 25 reservoirs that comprise the Sistema Interconectado Nacional (SIN) hydroelectric storage. Reservoirs are grouped in five dispatch regions: Antioquia (Amani, Miraflores, Peñol, Playas, Porce II, Porce III, Punchiná, Riogrande2, San Lorenzo, Troneras), Caribe (Urrá I), Centro (Agregado Bogotá, Betania, El Quimbo, Muna, Prado, Topocoro), Oriente (Chuza, Esmeralda, Guavio) and Valle (Altoanchicaya, Calima I, Salvajina). Aggregate, regional and per-reservoir series are all published. Notably, Chuza (the main Chingaza system reservoir supplying Bogotá's drinking water) is included, narrowing the energy-vs-supply gap. Coverage remains weak for purely irrigation reservoirs, smaller municipal supply infrastructure, and for the San Rafael complement of Chingaza, which is reported by EAAB separately and not integrated in XM's daily series.

Two complementary public APIs serve XM/SIN reservoir data: (a) the SIMEM PublicData endpoint at simem.co/backend-files/api/PublicData (datasetid + startDate + endDate parameters, no authentication, returns CSV/JSON/XML); and (b) the SINERGOX-XM REST API exposed via servapibi.xm.com.co, also unauthenticated. A maintained Python wrapper, pydataxm v0.3.14 on PyPI, returns pandas DataFrames; supported metrics include 'VolumenUtilDiarioPorEmbalse' and 'CapacidadUtilPorEmbalse', filterable per reservoir via 'ListadoEmbalse'. The API enforces a 30-day window per call for hourly/daily metrics (731 days for monthly), requiring pagination loops for multi-year pulls. IDEAM's DHIME portal exposes hydrometeorological series through an interactive UI only, with no machine-readable API, dragging the overall accessibility score down from what XM alone would warrant.

Methodology v1.3.0 (capacity-weighted), conservative denominator. Reported Colombian national hydroelectric storage capacity per XM/IDEAM cross-reference is approximately 22,000 hm³, dominated by Topocoro/Hidrosogamoso (~4,800 hm³), El Quimbo (~3,205 hm³), Guavio (~950 hm³), Betania (~1,971 hm³), Peñol (~1,200 hm³), Punchiná (~660 hm³), Salvajina, Calima I, Riogrande II, Esmeralda (Chivor), Miraflores, Playas, Porce II/III, Chuza (Chingaza supply), plus a tail of smaller water-supply and irrigation reservoirs. Applying a +10% conservative uplift for non-SIN municipal supply reservoirs (Medellín's San Rafael, EAAB Chingaza secondary, EPM and EMCALI supply impoundments), CORPOICA/INCODER irrigation reservoirs in the Caribbean and Andean regions, plus regional autonomous corporation (CAR) reservoirs that fall outside XM's dispatch perimeter, the realistic denominator is approximately 24,200 hm³. Covered capacity through XM's published SIN series (25 reservoirs including Chuza for Bogotá supply) is approximately 17,000 hm³. Coverage = round(100 × 17,000 / 24,200) = 70. The residual gap is San Rafael (the Chingaza system's secondary reservoir), municipal supply reservoirs for Medellín and Cali, and irrigation infrastructure for which IDEAM does not maintain a curated reservoir-fill dataset.

Honest reassessment downward from the first pass. The SIMEM portal was launched in September 2023, so the genuinely API-accessible time series via simem.co covers only about 2.5 years as of mid-2026. The legacy SINERGOX archive does host older hydrology reports back to roughly the early 2000s, but those are published as PDFs ('Boletín de Hidrología', annual 'Informe Anual'), not as a clean programmatic series. The SINERGOX-XM REST API itself returns daily reservoir data going back further than SIMEM but pre-2010 coverage is uneven across reservoirs. IDEAM's DHIME provides validated hydrological station series back to the 1960s–1980s, but as river-stage and discharge data, not as reservoir fill percentages. The previous evaluation of 72 conflated 'data exists somewhere' with 'data is accessible in usable form'; a more accurate read of programmatic, per-reservoir historical depth is roughly 10–15 usable years, with the cleanest series only ~3 years deep.

XM publishes daily useful volume data for all SIN reservoirs as part of its dispatch operations. Daily figures are posted on xm.com.co, distributed via official communiqués (including the elevated communications activity during the 2023–2024 El Niño drought and the renewed alerts in May 2026 as a fresh El Niño probability reached 82%), and accessible programmatically through the SIMEM and SINERGOX APIs. Mainstream media (El Tiempo, La República, Valora Analitik) republish the daily series. This is among the highest update frequencies achievable for reservoir monitoring — daily publication is systematic and driven by regulatory obligation: CREG Resolution 025/1995 mandates daily hydrological reporting by generators, formalized further in CNO operational agreements.

The methodology for computing useful volume is publicly described: water stored above the Technical Minimum Level (Nivel Mínimo Técnico), aggregated and expressed both as a percentage of total useful capacity and converted to GWh using XM's energy-equivalence coefficients. CREG Resolution 025/1995 and subsequent CNO agreements (notably Acuerdo CNO 512/2010) formalize reporting obligations. Capacity curves (cota–volumen) must be updated after each bathymetric survey. However, detailed station-level metadata — exact sensor methods, individual measurement uncertainty, the date of the last sedimentation survey per reservoir, post-survey capacity revisions — is not consolidated in a single public technical reference. IDEAM maintains methodological documentation for its hydrological network but it is scattered across agency reports. There is no equivalent of the USGS station-page metadata model.

All primary data portals — xm.com.co, simem.co, sinergox.xm.com.co, paratec.xm.com.co and ideam.gov.co — are exclusively in Spanish. There is no English-language interface anywhere in the chain. The GitHub repository for the API_XM toolkit (EquipoAnaliticaXM) is also Spanish-only in its documentation, although the JSON API field names are language-neutral and an English-speaking developer with technical Spanish reading ability can reverse-engineer the structure. The IEA 2023 Energy Policy Review for Colombia and some academic publications provide English-language secondary context, but the data source infrastructure itself offers zero English localization — lower than Brazil (where some ONS metadata is partially bilingual) and lower than Spain (which at least exposes confederation websites with partial English summaries).