Long Period Average (LPA) of Rainfall

The Long Period Average (LPA) of rainfall is the statistical baseline against which the India Meteorological Department (IMD) measures every monsoon season. It is the arithmetic mean of rainfall recorded across a fixed reference window of several decades, computed both for the all-India southwest monsoon (June–September) and for individual meteorological subdivisions, states, and districts. The concept originates in the IMD's long-range forecasting practice, formalised in the modern era after Gilbert Walker's early twentieth-century work on monsoon correlations and refined through successive statistical and dynamical models. The LPA exists because rainfall in a tropical monsoon regime is highly variable year to year; a single season's total means little without a stable normal to compare it against. By convention the IMD periodically updates the reference period—the all-India southwest monsoon LPA was long fixed at 88 cm based on the 1951–2000 average, and the IMD has updated its baselines to more recent climatological windows (such as 1961–2010 and 1971–2020) to reflect shifting rainfall patterns.

Procedurally, the LPA is constructed by aggregating daily and monthly rainfall observations from the IMD's network of surface stations, automatic weather stations, and increasingly satellite- and radar-derived estimates across the chosen multi-decade period. For each grid or subdivision the seasonal totals are averaged to yield a single normal value. The IMD then expresses any forecast or observed season as a percentage of LPA rather than in absolute millimetres, because a percentage communicates deviation more intuitively to planners and the public. A forecast that the monsoon will deliver "98 per cent of LPA" therefore states that the season is expected to fall two percentage points below the long-period normal. Each operational forecast also carries a stated model error, conventionally ±4 or ±5 per cent, reflecting the inherent uncertainty of seasonal prediction.

The IMD classifies each season into five categories defined as bands around the LPA. Rainfall between 96 and 104 per cent of LPA is termed normal; 90–96 per cent is below normal; less than 90 per cent is deficient (the threshold associated with drought); 104–110 per cent is above normal; and more than 110 per cent is excess. These thresholds are applied at the all-India scale and, with adjustments, to regional forecasts. The IMD issues its first long-range forecast in April and updates it in late May or early June before onset, supplementing the headline number with probabilistic category forecasts and, since the introduction of dynamical and coupled ocean-atmosphere models, spatial maps of expected distribution. The agency also tracks the monsoon's progress against the LPA cumulatively as the season unfolds.

In contemporary practice the LPA framework anchors the most consequential annual economic forecast in India. When the IMD in New Delhi announces in April that the southwest monsoon will be, for instance, 103 per cent of LPA, the figure propagates immediately to the Ministry of Agriculture and Farmers' Welfare, the Reserve Bank of India's inflation projections, and commodity markets. Drought-relief mechanisms under the National Disaster Management framework are triggered partly by district-level rainfall falling below LPA thresholds. The deficient monsoons of 2009, 2014, and 2015—each well below 90 per cent of LPA—prompted central drought declarations, while the 2019 and 2020 seasons exceeded LPA. Skymet, a private forecaster, issues parallel LPA-based predictions that are frequently contrasted with the IMD's in the press.

The LPA must be distinguished from several adjacent measures. It is not the same as the normal onset date of the monsoon, which is a calendar marker rather than a rainfall quantity. It differs from a rainfall anomaly, which is the signed departure of an observation from the LPA expressed in millimetres or percentage points. It is also distinct from the Standardised Precipitation Index, a statistical drought metric that normalises rainfall by its variability rather than by a simple mean. Crucially, a season at exactly 100 per cent of LPA at the national scale can still conceal severe regional deficits, because the all-India LPA aggregates wet and dry subdivisions; spatial and temporal distribution matter as much as the headline percentage.

The framework attracts recurring controversy on two fronts. First, the periodic revision of the reference period changes the LPA value itself, so a fixed absolute rainfall total can shift category between baselines—an updated, lower LPA can make the same rainfall appear closer to normal, which critics argue masks long-term climatic drying in some regions. Second, the all-India percentage is a blunt communicator: the headline "normal monsoon" reassures markets even when temporal clustering of rain or prolonged dry spells damages crops. Climate change has widened intra-seasonal variability, increasing the frequency of short intense spells and dry breaks that the seasonal LPA figure cannot capture. The IMD has responded by emphasising monthly and spatial forecasts and adopting multi-model ensemble and dynamical prediction systems alongside the legacy statistical approach.

For the working practitioner—the agriculture desk officer, the central-bank economist, the disaster-management coordinator, or the UPSC aspirant studying Indian geography—the LPA is indispensable but must be read with discipline. It is the single most cited number in India's annual climate-economic cycle, yet its meaning depends entirely on the reference period in use, the spatial scale of application, and the distribution behind the aggregate. A competent analyst never quotes a percentage of LPA without asking which baseline period it rests on, whether it is national or subdivisional, and how the rain was distributed across the season. Understood this way, the LPA is less a verdict than a starting point for interpreting the monsoon's economic and humanitarian consequences.