Temporal patterns in radar‐observed convective cell development during the 2016 monsoon onset

Monsoon onset is highly variable year-to-year and difficult to predict. Here, Doppler radar data shed light on convective development in the days weeks around 2016 monsoon at multiple sites across India. We consider this with relation large-scale dynamic thermodynamic changes associated onset. The timing of South Asian summer (hereafter referred as monsoon) notoriously forecast (e.g. Rao et al., 2019), a standard deviation 8 (Pai Nair, 2009). This interannual variability arrival progression has significant ramifications for communities India, who rely rains alleviate pre-monsoon heat, well agricultural production (Gadgil Gadgil, 2006). Climatologically, arrives southern India (Kerala) 1 June, before moving non-steady motion toward northwest, reaching far northwest border by 15 July. Parker al. (2016) noted climatologically gradual withdrawal dry air 400–700hPa, originating desert cumulus congestus clouds moistening mid-levels during advance, thus resulting formation more convection. In way, convection successive regions advance mid-level drives forwards, perpendicular low-level winds. Volonté (2020) found competition between retreating subtropical advancing moist tropical (as detailed 2016) be non-linear process, modulated through intraseasonal (active break periods; Rajeevan 2010) occurrence synoptic-scale features such cyclonic circulation over Arabian Sea. results periods hiatus forward jumps monsoon, non-steadiness further consequences predictability. Furthermore, Kerala tip substantial variability: since 1970 date stated Meteorological Department (IMD) fluctuated from early 19 May 1990 late 18 June 1972. combined aforementioned makes an inherently unpredictable phenomenon Whilst cloud physical processes involved are crucial components their representation weather climate models remains challenge, bias Indian subcontinent many (Willetts 2017). Parametrisations dominant source error global (Sherwood 2014) suffer historical undersampling observations help constrain these models. using IMD 12 country, we assess temporal patterns local fields 2016, compared increase rainfall. study aims build spatial how presents itself terms development. order analyse any regional differences, radars (Table 1), forming part network. Sixteen season were made available INCOMPASS (Turner 2020), but two did not have fully capturing A missing amounts region, so analysed study. remaining all S-band exception New Delhi (C-band), sampling 10 min volume scans. All similar beamwidths (0.93°–1.00°) elevation angles (see Table or identical scanning strategies. There some disparity gate size different radars, varying 250m 1000m Patna Karaikal respectively. calibrated prior analysis clutter reflectivity Global Precipitation Measurement (GPM) overpasses following methodology Warren (2018) Louf (2019). height reconstruct 3-D Cartesian grid each 2 16km above surface. method described concisely here reasons space, procedure that Kumar (2013), where it detail. Two kilometres ground contamination safely below freezing level (brightband) region. maximum chosen maximise range echo-top retrievals: upper edge beam angle 21° reaches 40km range. Any penetrating set 16.5km may represent deeper than this. horizontal resolution 1km × matches broadest within 100km vertical spacing 0.5km. profile calculated point interpolating separate Once obtained, Steiner (1995) convective/stratiform classification implemented reconstructed 2km identify pixels. region then connected pixels 4-connectivity. pixel assumed throughout profile, similarly (2013). Each scan given (ETH) based 10dBZ level. threshold low possible minimum sensitivity 1). Finally, cell-top (CTH) determined ETH cell greater 4km2 considered analysis. define cells CTH ≥ 8km deep cells, representing most potent Deep high reflectivities (typically excess 30dBZ, shown). also prescribe those 5km < less 30dBZ (light rain), approximate only single due availability, representative year analysis, 97% average all-India rainfall (7 later climatological average, one deviation; Pai definition particular location trivial. Difficulty stems need avoid false onsets. For example, southeastern saw record-breaking Cyclonic Storm Roanu, was regime. However, storms can act northern limit (Krishnamurti 1981). misconception think there no rain consistent after. rain-shadow (such lee Western Ghat mountains) will often experience long spells arrival. Due nature day-to-day complex task objectively Purely dynamical methods make sure synoptic regime (i.e. westerly winds lower troposphere easterly troposphere) little actual Statistical prone onsets result. historically used blend measured Recently, new objective criteria released 2020). uses gridded daily (Rajeevan 1° pixel, ensure day onset, stricter certain depending meteorology. Further details provided (2020). investigation, even especially Therefore, site, compare against isochrones . requirement consecutive 2.5mm least 60% surrounding stations, depth westerlies reach 600hPa satellite-derived outgoing longwave radiation (OLR) 200Wm−2 grid-box. takes into consideration subjective manner determine first which exact behind line taken site. site shown 1. Firstly, its discuss advance. Figure shows totals GPM Integrated Multi-satellitE Retrievals (IMERG) product (Huffman 2015), alongside 850hPa wind vectors ERA5 (ECMWF re-analysis; Hersbach averaged pentads (grouped 5-day periods) mid-May mid-July 2016. resolutions IMERG 0.1° 0.25° yet established Low-level land, advection, weak magnitude. Westerlies south gradually spread northwards, encompassing end (Figure 1(c)). Over land (except India), weaken again 1(d)) continue strengthen Sea (west India). As result, begins offshore western coast off coast. brings week June. this, period develops. central zone (18–27°N, 73–82°E) comparing 2010 provides details). Break decreased (and vice versa active periods). These events instrumental delaying northwestwards 2010). movement Figures 1(e) (f)) clear strength onshore (Fletcher Bay Bengal (east northward-pushing wave packet Boreal Summer Intraseasonal Oscillation (BSISO) phases 3 (Kikuchi 2012) much northeastern mid-June 1(g)), depression, 1(g) (h), depression visibly tracks westwards field, remainder July, Patiala push sufficiently northwards. Note earlier followed drier still northwesterly component, defined regions. By 1(l)), almost peak, persistent precipitation country. One major feature Asia establishment jet (TEJ). TEJ thermal forms spring developing north-to-south temperature gradient Tibetan Plateau equator. It continues until responsible usual westward passage depressions (h)), forcing (Parker 2016). Studying relative humidity moisture flux 2), general note retreat northwestern (2016). 2(c)), advects back delays Kerala. Into final pentad (near peak increasingly wedge spatially transient season, sometimes neighbouring Pakistan, pushing southeastward 2(h) (k)). Intriguingly seems bearing 2(j), once encroaching 2(j)), do north time. illustrates complexities predicting competing effects always displaying anti-correlation Moisture elucidate eastward convergence occur both sides (white dashed line), ocean. (2016), therefore moisture-laden alone cannot explain expect location. raises importance additional when defining now turn our attention 1, measurements 2010), evolve running means area (CTH 8km) type (5 ≤ 8km). black represents calculate Pearson correlation coefficients congestus/deep area, labelled positive (0.26 0.94 Mumbai). usually larger surface intensity. (Kolkata Delhi) show weaker correlations 0.5. Interestingly, relatively higher area. examining 3(e) (k) closely, poor stem large amount reflected time series Kolkata 25 8–15 July). could because covered congestus-type significantly sites, positively correlated 0.92 Mumbai likely predominance shallow precipitating (Utsav follows coverage smaller coverage. suggests resolution, approximately simultaneous levels. contrast progressive deepening timescale, might infer framework troposphere, allowing form recall OLR cold tops convection) observed precipitation. expect, few IMD-defined 3. immediately obvious rainfall, making purely observations-based definitions difficult. Some Mumbai, Nagpur Bhopal west pronounced pre- post-monsoon. Other locations picture if composite several seasons considered, become meaningful. Nonetheless, should too literally, owing high-frequency initiation. being hybrid neither properties isolation linked date. Karaikal, Vishakhapatnam, Bhopal, occurring BSISO frequency after Southeastern shadow near-zero contrast, experiences plentiful interesting abnormality presence (pre-onset) (Figures 3(i) (j)) (Mumbai signal though quite evident). other does result satellite able inform seemingly non-precipitating authors gratefully acknowledge agreement India's Ministry Earth Sciences providing under project, funded Mission. work supported SCENARIO, grant number NE/L002566/1. declare conflict interest. Thanks anonymous reviewers whose constructive comments greatly improved overall quality paper.