Cold pools over the Netherlands: A statistical study from tower and radar observations

Abstract

We provide a detailed analysis of convectively generated cold pools (CPs) over flat midlatitude land, combining ten‐year high‐frequency time series of measurements at several heights available from the 213‐m tower observatory at Cabauw, the Netherlands, with a collocated 2D radar rainfall dataset. This combination of data allows us to relate observations of the CP's temporal and vertical structure to the properties of each CP's parent rain cell, which we identify by rain‐cell tracking. Using a new detection method, based on the anomalies of both the vertically averaged wind and the temperature, we monitor the arrival and passing of 189 CPs during ten summers (2010–2019). The time series show a clear signature of vortex‐like motion along the leading CP edge in the vertical and horizontal wind measurements. The arrival of CP gust fronts is characterized by a steep decrease in both temperature and moisture, with a recovery time of approximately two hours. We see no evidence of moisture rings on the gust front edge, and therefore no indications for thermodynamic convective triggering. From the tower data, we obtain a median CP temperature drop of Tdrop≈−2.9 K and a height‐averaged horizontal wind anomaly of Δumax≈ 4.4 m·s−1. Relating the individual CP's horizontal wind anomalies and temperature drops, we confirm the validity of the theoretical density current relationship, Δumax∝Tdrop1/2. We further propose a simple statistical model to relate the CP strength defined by Tdrop to the environmental properties influencing the CP: rain intensity and lower boundary‐layer saturation. A multivariate linear regression suggests a 1 K colder CP for a 4 mm·hr−1 more intense rain cell (instantaneous area‐averaged rain intensity) or for a 2.5 K larger pre‐CP dew‐point depression.

Assessing Memory in Convection Schemes Using Idealized Tests

“…For the UW‐ org scheme, under FixMacro conditions we have observed a monotonic relationship between P and org , as presented in Section 3.2. For simplicity, and motivated by findings of previous studies (e.g., Kirsch et al., 2021; Kruse et al., 2022), we assume a linear approximation: $$$P=\beta \ org,$$$ where β is the proportionality factor. Equation , describing the UW‐ org scheme, is roughly equivalent to Equation of the PP model, with V ≡ org , α damp ≡ 1/ τ org , and α vp P ≡ evap 2 org · E , where E is the mass‐weighted vertical integral of rain evaporation rate and is proportional to (1 − RH) multiplied by the square root of P (Equation A8 in Park & Bretherton, 2009).…”

“…For the UW-org scheme, under FixMacro conditions we have observed a monotonic relationship between P and org, as presented in Section 3.2. For simplicity, and motivated by findings of previous studies (e.g., Kirsch et al, 2021;Kruse et al, 2022), we assume a linear approximation:…”

On the Sensitivity of Convective Cold Pools to Mesh Resolution

“…Several studies have carefully analyzed the effects of numerical model resolution of CPs (Bryan et al., 2003; Grant & van den Heever, 2016; Hirt et al., 2020; Huang et al., 2018; Moseley et al., 2020; Pressel et al., 2015; Straka et al., 1993; Zuidema et al., 2017), yet the spatial scales involved at the CP gust fronts are often finer than 100 m, as is evident when inspecting the spatial gradient in vertical velocities in observational studies (Kruse et al., 2022). For example, recent results by Grant and van den Heever (2016) advised that simulations should use grid spacings of at most 100 m horizontally and no more than 50 m vertically.…”

U‐Net Segmentation for the Detection of Convective Cold Pools From Cloud and Rainfall Fields

“…At Earth’s surface, CPs spread horizontally as density currents (Drager & van den Heever, 2017; Droegemeier & Wilhelmson, 1985; Zuidema et al., 2017). While expanding radially along the surface, CPs can be characterized as consisting of (a) a head, which can measure hundreds to several thousand meters vertically and (b) a shallower interior, which is separated from the head by a wake region (Benjamin, 1968; Cafaro & Rooney, 2018; Droegemeier & Wilhelmson, 1987; Fiévet et al., 2023; Kneller et al., 1999; Kruse et al., 2022; Markowski & Richardson, 2011; Meyer & Haerter, 2020; Simpson, 1980).…”