Large-scale vertical velocity, diabatic heating and drying profiles associated with seasonal and diurnal variations of convective systems observed in the GoAmazon2014/5 experiment

This study describes the characteristics of largescale vertical velocity, apparent heating source (Q1) and apparent moisture sink (Q2) profiles associated with seasonal and diurnal variations of convective systems observed during the two intensive operational periods (IOPs) that were conducted from 15 February to 26 March 2014 (wet season) and from 1 September to 10 October 2014 (dry season) near Manaus, Brazil, during the Green Ocean Amazon (GoAmazon2014/5) experiment. The derived large-scale fields have large diurnal variations according to convective activity in the GoAmazon region and the morning profiles show distinct differences between the dry and wet seasons. In the wet season, propagating convective systems originating far from the GoAmazon region are often seen in the early morning, while in the dry season they are rarely observed. Afternoon convective systems due to solar heating are frequently seen in both seasons. Accordingly, in the morning, there is strong upward motion and associated heating and drying throughout the entire troposphere in the wet season, which is limited to lower levels in the dry season. In the afternoon, both seasons exhibit weak heating and strong moistening in the boundary layer related to the vertical convergence of eddy fluxes. A set of case studies of three typical types of convective systems occurring in Amazonia – i.e., locally occurring systems, coastal-occurring systems and basin-occurring systems – is also conducted to investigate the variability of the large-scale environment with different types of convective systems.


5
2014; Burleyson et al., 2016). Schumacher et al. (2007) examined the diurnal cycle of the large-78 scale Q 1 budget in the southwest Amazon during LBA, but used only two profiles per day, which 79 do not capture the rapidly changing environment. In addition, the diurnal cycle over the highly 80 deforested southwest Amazon is not necessarily representative of the more pristine central 81 Amazonian rainforest.

82
In this study we use data collected from the comprehensive GoAmazon2014/5 field 83 campaign to examine the seasonal and diurnal variations of the large-scale vertical velocity and 84 heat and moisture budgets associated with the convective systems that occur in central Amazonia. 85 Section 2 provides details of the data and method used to derive the large-scale profiles for the 86 GoAmazon2014/5 experiment. Section 3 describes the synoptic conditions for the two IOPs.

92
Due to the lack of an appropriate sounding array to capture the divergence and advection 93 fields in the analysis domain, the large-scale vertical velocity and budgets analyzed in this study 94 were derived by using, as a first guess, the European Centre for Medium-Range Weather 95 Forecasts (ECMWF) analysis data that are subsequently constrained with surface and top of 96 atmosphere (TOA) observations. The upper-level fields from ECMWF analysis data are 97 adjusted to conserve the vertical integration of mass, moisture and dry static energy through a 98 constrained variational analysis technique described in Zhang and Lin (1997) and Zhang et al. 99 Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-644, 2016 Manuscript under review for journal Atmos. Chem. Phys. (CAPPI) at 2.5 km above ground was used to generate the rain rate products using a single Z-R Mobile Facility site near Manacapuru (3.213°S, 60.598°W; "ARM site" in Figure 1).

119
Observations of latent and sensible heat fluxes at two other Brazilian research sites -K34 120 ("FLUXNET-BR Ma2" in Figure 1) and the Amazon Tall Tower Observatory ("ATTO Tower" 121 in Figure 1) -are also used. Because of the limited number of surface sites, it is challenging to 122 Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-644, 2016 Manuscript under review for journal Atmos. Cressman's objective analysis method (Cressman, 1959) (Martin et al., 2016). The top three rows of Figure 3 show the domain-averaged zonal (u) wind, 140 meridional (v) wind, and relative humidity relative to liquid water, from the adjusted ECMWF

166
We also define rainy (black dotted lines) and non-rain periods (gray lines) using a threshold of 167 0.2 mm hr -1 . A value of 0.2 mm hr -1 rather than 0 mm hr -1 is used because in some cases ground systems, respectively, which will be discussed in Section 5.

175
The non-rain vertical velocity profiles are relatively weak, with downward motion 176 dominating in the upper troposphere during both dry and wet seasons. The rainy vertical 177 velocity profiles show strong upward motion throughout the troposphere during both IOPs, but 178 the level of maximum upward motion is different. The upward motion during the rainy period of 179 IOP1 has a broad peak structure from ~700 to 300 hPa with the maximum at ~350 hPa. The IOP2 rainy period also has a broad peak but the maximum is at a much lower level (~550 hPa) 185 than in IOP1. Because the frequency of the rainy period is higher in IOP1 than in IOP2, the IOP-186 mean upward motion is stronger during IOP1 but weaker and limited to the lower troposphere 187 during IOP2. As discussed in the next section, the difference in morning precipitation systems 1984; Lau and Peng, 1987;Puri, 1987;Hack and Schubert, 1990). (at 500 and 750 hPa) are higher than those in IOP2 (at 650 and 800 hPa). The double peak 213 features of Q 1 and Q 2 are likely due to different physical processes. For Q 1 , the local minimum 214 usually occurs near the melting level (~600 hPa), indicating latent cooling due to ice melting.

215
Because the melting level is nearly constant in the tropics, the local minimums of Q 1 are more or 216 less at the same level as seen in other tropical field campaigns (e.g. Schumacher et al., 2008;Xie 217 et al., 2010a). For Q 2 , the double-peak structure is the combined effect of convective (lower 218 peak) and stratiform (higher peak) rain production (Lin and Johnson, 1996). The peak levels for 219 stratiform and convective clouds may vary in different locations and times such as in the two 220 IOPs in this study. precipitation occurs in the afternoon. The magnitude of afternoon precipitation in IOP2 is just 227 slightly smaller than that in IOP1, but the magnitude of morning precipitation in IOP2 is 228 significantly lower than that in IOP1, indicating that the differences between dry and wet seasons 229 are mainly due to the morning precipitation events. The surface CAPE has similar magnitudes in 230 the daytime during IOP1 and IOP2, but in the early morning it rises later and slower during IOP1 231 than during IOP2, probably because early morning precipitation during IOP1 has released 232 atmospheric instability. The surface CIN is typically small, especially during IOP1, which is due 233 to the high surface relative humidity over the Amazon rainforest.

234
The diurnal cycles of cloud frequency, large-scale vertical velocity, Q 1 , Q 2 and 1 2 QQ  235 for IOP1 (left) and IOP2 (right) are shown in Figure 8. Derived from Eq. (1) and (2), where v h s L q  is the moist static energy. With the phase change of water vapor cancelled, The early afternoon upward motion peaks at the upper troposphere and extends above 100 hPa.

243
Accordingly, clouds are mainly seen between 800 and 500 hPa in the early morning but  Table 1 gives the number of each type of precipitation system observed during the two IOPs, forcing data from the three-dimensional constrained variational analysis recently developed by 382 Tang and Zhang (2015) and Tang et al. (2016). This will be the subject of a future study.   relative humidity, (f) vertical velocity, (g) Q 1 and (h) Q 2 for the LOS case. The black lines are zero-lines. 603 The shaded and white areas in (b) indicate nightime and daytime. 604 relative humidity, (f) vertical velocity, (g) Q 1 and (h) Q 2 for the COS case. The black lines are zero-lines. 606 The shaded and white areas in (b) indicate nightime and daytime. 607 The shaded and white areas in (b) indicate nightime and daytime. 610 Atmos. Chem. Phys. Discuss., doi:10.5194/acp-2016-644, 2016 Manuscript under review for journal Atmos. Chem. Phys.