1 | SUBROUTINE lpm_droplet_condensation |
---|
2 | |
---|
3 | !------------------------------------------------------------------------------! |
---|
4 | ! Current revisions: |
---|
5 | ! ------------------ |
---|
6 | ! |
---|
7 | ! |
---|
8 | ! Former revisions: |
---|
9 | ! ----------------- |
---|
10 | ! $Id: lpm_droplet_condensation.f90 850 2012-03-15 12:09:25Z letzel $ |
---|
11 | ! |
---|
12 | ! 849 2012-03-15 10:35:09Z raasch |
---|
13 | ! initial revision (former part of advec_particles) |
---|
14 | ! |
---|
15 | ! |
---|
16 | ! Description: |
---|
17 | ! ------------ |
---|
18 | ! Calculates change in droplet radius by condensation/evaporation, using |
---|
19 | ! either an analytic formula or by numerically integrating the radius growth |
---|
20 | ! equation including curvature and solution effects using Rosenbrocks method |
---|
21 | ! (see Numerical recipes in FORTRAN, 2nd edition, p. 731). |
---|
22 | ! The analytical formula and growth equation follow those given in |
---|
23 | ! Rogers and Yau (A short course in cloud physics, 3rd edition, p. 102/103). |
---|
24 | !------------------------------------------------------------------------------! |
---|
25 | |
---|
26 | USE arrays_3d |
---|
27 | USE cloud_parameters |
---|
28 | USE constants |
---|
29 | USE control_parameters |
---|
30 | USE cpulog |
---|
31 | USE grid_variables |
---|
32 | USE interfaces |
---|
33 | USE lpm_collision_kernels_mod |
---|
34 | USE particle_attributes |
---|
35 | |
---|
36 | IMPLICIT NONE |
---|
37 | |
---|
38 | INTEGER :: i, internal_timestep_count, j, jtry, k, n |
---|
39 | |
---|
40 | INTEGER, PARAMETER :: maxtry = 40 |
---|
41 | |
---|
42 | REAL :: aa, afactor, arg, bb, cc, dd, ddenom, delta_r, drdt, drdt_ini, & |
---|
43 | drdt_m, dt_ros, dt_ros_last, dt_ros_next, dt_ros_sum, & |
---|
44 | dt_ros_sum_ini, d2rdtdr, d2rdt2, errmax, err_ros, g1, g2, g3, g4, & |
---|
45 | e_a, e_s, gg, new_r, p_int, pt_int, pt_int_l, pt_int_u, q_int, & |
---|
46 | q_int_l, q_int_u, ql_int, ql_int_l, ql_int_u, r_ros, r_ros_ini, & |
---|
47 | sigma, t_int, x, y |
---|
48 | |
---|
49 | ! |
---|
50 | !-- Parameters for Rosenbrock method |
---|
51 | REAL, PARAMETER :: a21 = 2.0, a31 = 48.0/25.0, a32 = 6.0/25.0, & |
---|
52 | a2x = 1.0, a3x = 3.0/5.0, b1 = 19.0/9.0, b2 = 0.5, & |
---|
53 | b3 = 25.0/108.0, b4 = 125.0/108.0, c21 = -8.0, & |
---|
54 | c31 = 372.0/25.0, c32 = 12.0/5.0, & |
---|
55 | c41 = -112.0/125.0, c42 = -54.0/125.0, & |
---|
56 | c43 = -2.0/5.0, c1x = 0.5, c2x= -3.0/2.0, & |
---|
57 | c3x = 121.0/50.0, c4x = 29.0/250.0, & |
---|
58 | errcon = 0.1296, e1 = 17.0/54.0, e2 = 7.0/36.0, & |
---|
59 | e3 = 0.0, e4 = 125.0/108.0, gam = 0.5, grow = 1.5, & |
---|
60 | pgrow = -0.25, pshrnk = -1.0/3.0, shrnk = 0.5 |
---|
61 | |
---|
62 | |
---|
63 | CALL cpu_log( log_point_s(42), 'lpm_droplet_condens', 'start' ) |
---|
64 | |
---|
65 | DO n = 1, number_of_particles |
---|
66 | ! |
---|
67 | !-- Interpolate temperature and humidity. |
---|
68 | !-- First determine left, south, and bottom index of the arrays. |
---|
69 | i = particles(n)%x * ddx |
---|
70 | j = particles(n)%y * ddy |
---|
71 | k = ( particles(n)%z + 0.5 * dz * atmos_ocean_sign ) / dz & |
---|
72 | + offset_ocean_nzt ! only exact if equidistant |
---|
73 | |
---|
74 | x = particles(n)%x - i * dx |
---|
75 | y = particles(n)%y - j * dy |
---|
76 | aa = x**2 + y**2 |
---|
77 | bb = ( dx - x )**2 + y**2 |
---|
78 | cc = x**2 + ( dy - y )**2 |
---|
79 | dd = ( dx - x )**2 + ( dy - y )**2 |
---|
80 | gg = aa + bb + cc + dd |
---|
81 | |
---|
82 | pt_int_l = ( ( gg - aa ) * pt(k,j,i) + ( gg - bb ) * pt(k,j,i+1) & |
---|
83 | + ( gg - cc ) * pt(k,j+1,i) + ( gg - dd ) * pt(k,j+1,i+1) & |
---|
84 | ) / ( 3.0 * gg ) |
---|
85 | |
---|
86 | pt_int_u = ( ( gg-aa ) * pt(k+1,j,i) + ( gg-bb ) * pt(k+1,j,i+1) & |
---|
87 | + ( gg-cc ) * pt(k+1,j+1,i) + ( gg-dd ) * pt(k+1,j+1,i+1) & |
---|
88 | ) / ( 3.0 * gg ) |
---|
89 | |
---|
90 | pt_int = pt_int_l + ( particles(n)%z - zu(k) ) / dz * & |
---|
91 | ( pt_int_u - pt_int_l ) |
---|
92 | |
---|
93 | q_int_l = ( ( gg - aa ) * q(k,j,i) + ( gg - bb ) * q(k,j,i+1) & |
---|
94 | + ( gg - cc ) * q(k,j+1,i) + ( gg - dd ) * q(k,j+1,i+1) & |
---|
95 | ) / ( 3.0 * gg ) |
---|
96 | |
---|
97 | q_int_u = ( ( gg-aa ) * q(k+1,j,i) + ( gg-bb ) * q(k+1,j,i+1) & |
---|
98 | + ( gg-cc ) * q(k+1,j+1,i) + ( gg-dd ) * q(k+1,j+1,i+1) & |
---|
99 | ) / ( 3.0 * gg ) |
---|
100 | |
---|
101 | q_int = q_int_l + ( particles(n)%z - zu(k) ) / dz * & |
---|
102 | ( q_int_u - q_int_l ) |
---|
103 | |
---|
104 | ql_int_l = ( ( gg - aa ) * ql(k,j,i) + ( gg - bb ) * ql(k,j,i+1) & |
---|
105 | + ( gg - cc ) * ql(k,j+1,i) + ( gg - dd ) * ql(k,j+1,i+1) & |
---|
106 | ) / ( 3.0 * gg ) |
---|
107 | |
---|
108 | ql_int_u = ( ( gg-aa ) * ql(k+1,j,i) + ( gg-bb ) * ql(k+1,j,i+1) & |
---|
109 | + ( gg-cc ) * ql(k+1,j+1,i) + ( gg-dd ) * ql(k+1,j+1,i+1) & |
---|
110 | ) / ( 3.0 * gg ) |
---|
111 | |
---|
112 | ql_int = ql_int_l + ( particles(n)%z - zu(k) ) / dz * & |
---|
113 | ( ql_int_u - ql_int_l ) |
---|
114 | |
---|
115 | ! |
---|
116 | !-- Calculate real temperature and saturation vapor pressure |
---|
117 | p_int = hyp(k) + ( particles(n)%z - zu(k) ) / dz * ( hyp(k+1)-hyp(k) ) |
---|
118 | t_int = pt_int * ( p_int / 100000.0 )**0.286 |
---|
119 | |
---|
120 | e_s = 611.0 * EXP( l_d_rv * ( 3.6609E-3 - 1.0 / t_int ) ) |
---|
121 | |
---|
122 | ! |
---|
123 | !-- Current vapor pressure |
---|
124 | e_a = q_int * p_int / ( 0.378 * q_int + 0.622 ) |
---|
125 | |
---|
126 | ! |
---|
127 | !-- Thermal conductivity for water (from Rogers and Yau, Table 7.1), |
---|
128 | !-- diffusivity for water vapor (after Hall und Pruppacher, 1976) |
---|
129 | thermal_conductivity_l = 7.94048E-05 * t_int + 0.00227011 |
---|
130 | diff_coeff_l = 0.211E-4 * ( t_int / 273.15 )**1.94 * & |
---|
131 | ( 101325.0 / p_int) |
---|
132 | ! |
---|
133 | !-- Change in radius by condensation/evaporation |
---|
134 | IF ( particles(n)%radius >= 1.0E-6 .OR. & |
---|
135 | .NOT. curvature_solution_effects ) THEN |
---|
136 | ! |
---|
137 | !-- Approximation for large radii, where curvature and solution |
---|
138 | !-- effects can be neglected |
---|
139 | arg = particles(n)%radius**2 + 2.0 * dt_3d * & |
---|
140 | ( e_a / e_s - 1.0 ) / & |
---|
141 | ( ( l_d_rv / t_int - 1.0 ) * l_v * rho_l / t_int / & |
---|
142 | thermal_conductivity_l + & |
---|
143 | rho_l * r_v * t_int / diff_coeff_l / e_s ) |
---|
144 | IF ( arg < 1.0E-16 ) THEN |
---|
145 | new_r = 1.0E-8 |
---|
146 | ELSE |
---|
147 | new_r = SQRT( arg ) |
---|
148 | ENDIF |
---|
149 | ENDIF |
---|
150 | |
---|
151 | IF ( curvature_solution_effects .AND. & |
---|
152 | ( ( particles(n)%radius < 1.0E-6 ) .OR. ( new_r < 1.0E-6 ) ) ) & |
---|
153 | THEN |
---|
154 | ! |
---|
155 | !-- Curvature and solutions effects are included in growth equation. |
---|
156 | !-- Change in Radius is calculated with 4th-order Rosenbrock method |
---|
157 | !-- for stiff o.d.e's with monitoring local truncation error to adjust |
---|
158 | !-- stepsize (see Numerical recipes in FORTRAN, 2nd edition, p. 731). |
---|
159 | !-- For larger radii the simple analytic method (see ELSE) gives |
---|
160 | !-- almost the same results. |
---|
161 | ! |
---|
162 | !-- Surface tension after (Straka, 2009) |
---|
163 | sigma = 0.0761 - 0.00155 * ( t_int - 273.15 ) |
---|
164 | |
---|
165 | r_ros = particles(n)%radius |
---|
166 | dt_ros_sum = 0.0 ! internal integrated time (s) |
---|
167 | internal_timestep_count = 0 |
---|
168 | |
---|
169 | ddenom = 1.0 / ( rho_l * r_v * t_int / ( e_s * diff_coeff_l ) + & |
---|
170 | ( l_v / ( r_v * t_int ) - 1.0 ) * & |
---|
171 | rho_l * l_v / ( thermal_conductivity_l * t_int )& |
---|
172 | ) |
---|
173 | |
---|
174 | afactor = 2.0 * sigma / ( rho_l * r_v * t_int ) |
---|
175 | |
---|
176 | IF ( particles(n)%rvar3 == -9999999.9 ) THEN |
---|
177 | ! |
---|
178 | !-- First particle timestep. Derivative has to be calculated. |
---|
179 | drdt_m = ddenom / r_ros * ( e_a / e_s - 1.0 - & |
---|
180 | afactor / r_ros + & |
---|
181 | bfactor / r_ros**3 ) |
---|
182 | ELSE |
---|
183 | ! |
---|
184 | !-- Take value from last PALM timestep |
---|
185 | drdt_m = particles(n)%rvar3 |
---|
186 | ENDIF |
---|
187 | ! |
---|
188 | !-- Take internal timestep values from the end of last PALM timestep |
---|
189 | dt_ros_last = particles(n)%rvar1 |
---|
190 | dt_ros_next = particles(n)%rvar2 |
---|
191 | ! |
---|
192 | !-- Internal timestep must not be larger than PALM timestep |
---|
193 | dt_ros = MIN( dt_ros_next, dt_3d ) |
---|
194 | ! |
---|
195 | !-- Integrate growth equation in time unless PALM timestep is reached |
---|
196 | DO WHILE ( dt_ros_sum < dt_3d ) |
---|
197 | |
---|
198 | internal_timestep_count = internal_timestep_count + 1 |
---|
199 | |
---|
200 | ! |
---|
201 | !-- Derivative at starting value |
---|
202 | drdt = ddenom / r_ros * ( e_a / e_s - 1.0 - afactor / r_ros + & |
---|
203 | bfactor / r_ros**3 ) |
---|
204 | drdt_ini = drdt |
---|
205 | dt_ros_sum_ini = dt_ros_sum |
---|
206 | r_ros_ini = r_ros |
---|
207 | |
---|
208 | ! |
---|
209 | !-- Calculate time derivative of dr/dt |
---|
210 | d2rdt2 = ( drdt - drdt_m ) / dt_ros_last |
---|
211 | |
---|
212 | ! |
---|
213 | !-- Calculate radial derivative of dr/dt |
---|
214 | d2rdtdr = ddenom * ( ( 1.0 - e_a/e_s ) / r_ros**2 + & |
---|
215 | 2.0 * afactor / r_ros**3 - & |
---|
216 | 4.0 * bfactor / r_ros**5 ) |
---|
217 | ! |
---|
218 | !-- Adjust stepsize unless required accuracy is reached |
---|
219 | DO jtry = 1, maxtry+1 |
---|
220 | |
---|
221 | IF ( jtry == maxtry+1 ) THEN |
---|
222 | message_string = 'maxtry > 40 in Rosenbrock method' |
---|
223 | CALL message( 'lpm_droplet_condensation', 'PA0347', 2, 2, & |
---|
224 | 0, 6, 0 ) |
---|
225 | ENDIF |
---|
226 | |
---|
227 | aa = 1.0 / ( gam * dt_ros ) - d2rdtdr |
---|
228 | g1 = ( drdt_ini + dt_ros * c1x * d2rdt2 ) / aa |
---|
229 | r_ros = r_ros_ini + a21 * g1 |
---|
230 | drdt = ddenom / r_ros * ( e_a / e_s - 1.0 - & |
---|
231 | afactor / r_ros + & |
---|
232 | bfactor / r_ros**3 ) |
---|
233 | |
---|
234 | g2 = ( drdt + dt_ros * c2x * d2rdt2 + c21 * g1 / dt_ros )& |
---|
235 | / aa |
---|
236 | r_ros = r_ros_ini + a31 * g1 + a32 * g2 |
---|
237 | drdt = ddenom / r_ros * ( e_a / e_s - 1.0 - & |
---|
238 | afactor / r_ros + & |
---|
239 | bfactor / r_ros**3 ) |
---|
240 | |
---|
241 | g3 = ( drdt + dt_ros * c3x * d2rdt2 + & |
---|
242 | ( c31 * g1 + c32 * g2 ) / dt_ros ) / aa |
---|
243 | g4 = ( drdt + dt_ros * c4x * d2rdt2 + & |
---|
244 | ( c41 * g1 + c42 * g2 + c43 * g3 ) / dt_ros ) / aa |
---|
245 | r_ros = r_ros_ini + b1 * g1 + b2 * g2 + b3 * g3 + b4 * g4 |
---|
246 | |
---|
247 | dt_ros_sum = dt_ros_sum_ini + dt_ros |
---|
248 | |
---|
249 | IF ( dt_ros_sum == dt_ros_sum_ini ) THEN |
---|
250 | message_string = 'zero stepsize in Rosenbrock method' |
---|
251 | CALL message( 'lpm_droplet_condensation', 'PA0348', 2, 2, & |
---|
252 | 0, 6, 0 ) |
---|
253 | ENDIF |
---|
254 | ! |
---|
255 | !-- Calculate error |
---|
256 | err_ros = e1*g1 + e2*g2 + e3*g3 + e4*g4 |
---|
257 | errmax = 0.0 |
---|
258 | errmax = MAX( errmax, ABS( err_ros / r_ros_ini ) ) / eps_ros |
---|
259 | ! |
---|
260 | !-- Leave loop if accuracy is sufficient, otherwise try again |
---|
261 | !-- with a reduced stepsize |
---|
262 | IF ( errmax <= 1.0 ) THEN |
---|
263 | EXIT |
---|
264 | ELSE |
---|
265 | dt_ros = SIGN( MAX( ABS( 0.9 * dt_ros * errmax**pshrnk ), & |
---|
266 | shrnk * ABS( dt_ros ) ), dt_ros ) |
---|
267 | ENDIF |
---|
268 | |
---|
269 | ENDDO ! loop for stepsize adjustment |
---|
270 | |
---|
271 | ! |
---|
272 | !-- Calculate next internal timestep |
---|
273 | IF ( errmax > errcon ) THEN |
---|
274 | dt_ros_next = 0.9 * dt_ros * errmax**pgrow |
---|
275 | ELSE |
---|
276 | dt_ros_next = grow * dt_ros |
---|
277 | ENDIF |
---|
278 | |
---|
279 | ! |
---|
280 | !-- Estimated timestep is reduced if the PALM time step is exceeded |
---|
281 | dt_ros_last = dt_ros |
---|
282 | IF ( ( dt_ros_next + dt_ros_sum ) >= dt_3d ) THEN |
---|
283 | dt_ros = dt_3d - dt_ros_sum |
---|
284 | ELSE |
---|
285 | dt_ros = dt_ros_next |
---|
286 | ENDIF |
---|
287 | |
---|
288 | drdt_m = drdt |
---|
289 | |
---|
290 | ENDDO |
---|
291 | ! |
---|
292 | !-- Store derivative and internal timestep values for next PALM step |
---|
293 | particles(n)%rvar1 = dt_ros_last |
---|
294 | particles(n)%rvar2 = dt_ros_next |
---|
295 | particles(n)%rvar3 = drdt_m |
---|
296 | |
---|
297 | new_r = r_ros |
---|
298 | ! |
---|
299 | !-- Radius should not fall below 1E-8 because Rosenbrock method may |
---|
300 | !-- lead to errors otherwise |
---|
301 | new_r = MAX( new_r, 1.0E-8 ) |
---|
302 | |
---|
303 | ENDIF |
---|
304 | |
---|
305 | delta_r = new_r - particles(n)%radius |
---|
306 | |
---|
307 | ! |
---|
308 | !-- Sum up the change in volume of liquid water for the respective grid |
---|
309 | !-- volume (this is needed later in lpm_calc_liquid_water_content for |
---|
310 | !-- calculating the release of latent heat) |
---|
311 | i = ( particles(n)%x + 0.5 * dx ) * ddx |
---|
312 | j = ( particles(n)%y + 0.5 * dy ) * ddy |
---|
313 | k = particles(n)%z / dz + 1 + offset_ocean_nzt_m1 |
---|
314 | ! only exact if equidistant |
---|
315 | |
---|
316 | ql_c(k,j,i) = ql_c(k,j,i) + particles(n)%weight_factor * & |
---|
317 | rho_l * 1.33333333 * pi * & |
---|
318 | ( new_r**3 - particles(n)%radius**3 ) / & |
---|
319 | ( rho_surface * dx * dy * dz ) |
---|
320 | IF ( ql_c(k,j,i) > 100.0 ) THEN |
---|
321 | WRITE( message_string, * ) 'k=',k,' j=',j,' i=',i, & |
---|
322 | ' ql_c=',ql_c(k,j,i), ' &part(',n,')%wf=', & |
---|
323 | particles(n)%weight_factor,' delta_r=',delta_r |
---|
324 | CALL message( 'lpm_droplet_condensation', 'PA0143', 2, 2, -1, 6, 1 ) |
---|
325 | ENDIF |
---|
326 | |
---|
327 | ! |
---|
328 | !-- Change the droplet radius |
---|
329 | IF ( ( new_r - particles(n)%radius ) < 0.0 .AND. new_r < 0.0 ) & |
---|
330 | THEN |
---|
331 | WRITE( message_string, * ) '#1 k=',k,' j=',j,' i=',i, & |
---|
332 | ' e_s=',e_s, ' e_a=',e_a,' t_int=',t_int, & |
---|
333 | ' &delta_r=',delta_r, & |
---|
334 | ' particle_radius=',particles(n)%radius |
---|
335 | CALL message( 'lpm_droplet_condensation', 'PA0144', 2, 2, -1, 6, 1 ) |
---|
336 | ENDIF |
---|
337 | |
---|
338 | ! |
---|
339 | !-- Sum up the total volume of liquid water (needed below for |
---|
340 | !-- re-calculating the weighting factors) |
---|
341 | ql_v(k,j,i) = ql_v(k,j,i) + particles(n)%weight_factor * new_r**3 |
---|
342 | |
---|
343 | particles(n)%radius = new_r |
---|
344 | |
---|
345 | ! |
---|
346 | !-- Determine radius class of the particle needed for collision |
---|
347 | IF ( ( hall_kernel .OR. wang_kernel ) .AND. use_kernel_tables ) & |
---|
348 | THEN |
---|
349 | particles(n)%class = ( LOG( new_r ) - rclass_lbound ) / & |
---|
350 | ( rclass_ubound - rclass_lbound ) * & |
---|
351 | radius_classes |
---|
352 | particles(n)%class = MIN( particles(n)%class, radius_classes ) |
---|
353 | particles(n)%class = MAX( particles(n)%class, 1 ) |
---|
354 | ENDIF |
---|
355 | |
---|
356 | ENDDO |
---|
357 | |
---|
358 | CALL cpu_log( log_point_s(42), 'lpm_droplet_condens', 'stop' ) |
---|
359 | |
---|
360 | |
---|
361 | END SUBROUTINE lpm_droplet_condensation |
---|