readme.txt

超强的大尺度水文模拟工具

# README.txt - Release notes
# $Id: README.txt,v 5.4.2.30 2004/10/04 20:52:30 vicadmin Exp $
#-----------------------------------------------------------------------

***** Description of changes from VIC 4.0.4 to VIC 4.0.5 *****
--------------------------------------------------------------------------------


New Features:
-------------

"-o" command-line option and display of run/compile-time options

	Files affected:
	cmd_proc.c, display_current_settings.c, get_global_param.c, global.h,
	vicNl.c, vicNl.h, vicNl_def.h

	Description:
	In VIC 4.0.5, if VERBOSE is TRUE, all compile-time options (from
	user_def.h) and run-time options (from your global parameter file) are
	displayed for you at the beginning of a model run.  If you are saving
	your model output to a log file, this information will be stored with
	your log, so you will have a record of the option settings that produced
	the results of that particular model run.
	In addition, added the "-o" option, to display current compile-time
	options.  One benefit of this option is that you can see what the
	options in user_def.h were set to when vicNl was compiled, without
	having to start a model run.  Since your version of user_def.h may have
	changed since you compiled vicNl, this is the most reliable way to see
	what these options are set to.

	Example:

	gen.hydro.washington.edu 41: vicNl -o

	***** VIC Version 4.0.5 - Current Model Settings *****

	COMPILE-TIME OPTIONS (set in user_def.h)
	----------------------------------------

	Output to Screen:
	VERBOSE                 TRUE

	Input Files:
	NO_REWIND               TRUE

	Output Files:
	LDAS_OUTPUT             FALSE
	LINK_DEBUG              FALSE
	OPTIMIZE                FALSE
	OUTPUT_FORCE            FALSE
	SAVE_STATE              FALSE

	Simulation Parameters:
	LOW_RES_MOIST           FALSE
	QUICK_FS                FALSE

	Maximum Array Sizes:
	MAX_BANDS               10
	MAX_FRONTS               3
	MAX_LAYERS               3
	MAX_NODES               18
	MAX_VEG                 12

	Snow Constants:
	NEW_SNOW_ALB            0.850000
	SNOW_ALB_ACCUM_A        0.940000
	SNOW_ALB_ACCUM_B        0.580000
	SNOW_ALB_THAW_A         0.820000
	SNOW_ALB_THAW_B         0.460000
	TraceSnow               0.030000

	Other Constants:
	LAI_WATER_FACTOR        0.200000
	LWAVE_COR               1.000000
	MAXIT_FE                25


Automatic recompilation on updates to *.h

	Files affected:
	Makefile

	Description:
	In VIC 4.0.4 and earlier, updating a .h file and recompiling VIC would
	not result in recompilation of files that depend on the .h file, unless
	a "make clean" command was issued first.  Now, if any .h files are
	updated, all dependent .c files are recompiled on the next "make".


Optional July average temperature field in soil parameter file

	Files affected:
	check_files.c, compute_treeline.c, initialize_atmos.c,
	initialize_global.c, read_soilparam.c, read_soilparam_arc.c, vicNl.c,
	vicNl.h, vicNl_def.h

	Description:
	In 4.0.4, when using the COMPUTE_TREELINE option, if the forcing files
	did not contain data for at least one July, the computed July average
	temperature would be undefined, resulting in incorrect location of the
	treeline.  In addition, runs covering different periods of the same
	forcing data would have different July average temperatures, causing the
	location of the treeline to change from run to run.  To allow more user
	control over the location of the treeline, added an optional field to
	the soil file, containing average July air temperature.  If the soil
	file contains 54 fields (or if the arc/info soil file list has 1 more
	line), this final entry is assumed to be the average July air
	temperature.  If COMPUTE_TREELINE is TRUE, this temperature is used in
	computing the treeline rather than calculating the average July
	temperature from the forcing data.  If COMPUTE_TREELINE is FALSE, this
	field is ignored.  If this field is not present in the soil file, vic
	behaves as in 4.0.4.


Bug Fixes:
----------

Spurious condensation at low temperatures

	Files affected:
	arno_evap.c

	Description:
	Changed logic of evap limit check to avoid creating spurious
	condensation.  In VIC 4.0.4 and earlier, whenever evaporation > (liquid
	moisture - residual moisture), evaporation would be set to (liquid
	moisture - residual moisture).  However, at low temperatures, when most
	or all soil moisture is frozen, liquid moisture < residual moisture,
	causing (liquid moisture - residual moisture) to be negative.  Any non-
	negative evap would be greater than this, resulting in evap getting set
	to (liquid moisture - residual moisture), which would be negative (i.e.
	condensation).  This artificially created condensation in whatever
	amount necessary to bring liquid moisture up to residual, causing 1)
	large latent heat flux, 2) incorrect surface temperatures, 3) occasional
	inability for calc_surf_energy_bal to converge in root_brent, and 4)
	spuriously high runoff and baseflow.  Now there is an added condition
	that liquid moisture > residual moisture for evap to be capped at
	(liquid moisture - residual moisture).

	NOTE: This fix results in lower runoff and baseflow in unvegetated areas
	with frozen soils, and may require recalibration of soil parameters.


Validation of initial soil moisture

	Files affected:
	read_soilparam.c, read_soilparam_arc.c

	Description:
	Added check to read_soilparam.c and read_soilparam_arc.c to make sure
	that wilting point is greater than residual moisture.  Changed lower
	limit on initial soil moisture to be residual moisture rather than
	wilting point.  Made validation statements clearer.  Validation does
	not occur if soil moisture will be read from a model state file.
	(found and fixed by Chunmei Zhu, Alan Hamlet, and Ted Bohn)

	NOTE: Soil parameter files containing Wpwp_FRACT and resid_moist
	such that Wpwp_FRACT < resid_moist / (1.0 - bulk_density/soil_density)
	will now cause VIC to exit with an error message.


Incorrect baseflow limits

	Files affected:
	runoff.c

	Description:
	In 4.0.4, runoff.c checked for the wrong bounds on baseflow, allowing
	baseflow to become negative when liquid soil moisture < residual
	moisture.  These bounds have been fixed in 4.0.5, as follows: baseflow
	is not allowed to exceed (liquid soil moisture - residual moisture); when
	baseflow < 0, baseflow is set to 0; when baseflow > 0 and the resulting
	soil moisture < residual moisture, water is taken out of baseflow and
	given to the soil as follows:
	  if baseflow > (residual moisture - soil moisture), then
	    baseflow -= (residual moisture - soil moisture);
	    soil moisture += (residual moisture - soil moisture);
	  else
	    soil moisture += baseflow;
	    baseflow = 0;

	NOTE: This fix may result in small changes in baseflow and evaporation.


Runs using initial state files starting at state file date rather than global
start date

	Files affected:
	check_state_file.c

	Description:
	In 4.0.4, check_state_file.c would increment the index of the forcing
	data array until it reached the record corresponding to the date
	stored in the state file.  This caused the simulation to start at the
	date at which the state file was saved rather than the start date
	specified in the global parameter file.  If the state file's date was
	earlier than the start date in the global parameter file, the index
	would be incremented until a segmentation fault occurred.  This has
	been fixed in 4.0.5 so that the start date in the global parameter
	file is always the start date of the simulation.  The date stored in
	the initial state file is ignored.

	NOTE: If you have been relying on the state file to dictate when your
	simulations start, this fix may require you to change your global
	parameter file so that STARTYEAR, STARTMONTH, etc. reflect the start
	date/time you want.


Negative incoming shortwave radiation at high latitudes

	Files affected:
	mtclim42_vic.c

	Description:
	In 4.0.4, when sub-daily shortwave radiation is estimated from daily
	min/max temperatures, negative values occasionally are calculated in
	coastal areas above the Arctic Circle in winter.  Now, if estimated sub-
	daily incident shortwave is negative, it is set to 0.0.


Undefined daily precipitation for deserts

	Files affected:
	mtclim42_vic.c

	Description:
	In 4.0.4, if a grid cell's annual precipitation (specified in the soil
	parameter file) is 0, then the adjusted daily precipitation calculated
	in mtclim42_vic.c ends up being undefined.  In 4.0.5 this has been
	fixed.

	More specifically, the MTCLIM 4.2 algorithm (which VIC uses for
	estimating sub-daily forcing values) was originally set up to expect a
	base precipitation as an input, and to translate this base precip into
	a site-specific precip via an adjustment function.  The adjustment
	function multiplies the base precip by the ratio of site_isohyet to
	base_isohyet to get site precipitation.  In calling the MTCLIM 4.2
	functions, VIC sets site_isohyet and base_isohyet to the grid cell's
	annual precipitation.  So this ratio should always = 1.  However,
	when annual precipitation is 0.0, this ratio is undefined.  So, the
	fix is to set the ratio to 1 when both site_isohyet and base_isohyet
	are 0 (or very small). (found by Liz Clark)


100% snow when air_temp = MAX_SNOW_TEMP

	Files affected:
	calc_rainonly.c

	Description:
	In 4.0.4, when air_temp = MAX_SNOW_TEMP, the portion of precipitation
	that is snow was set to 100%.  This has been fixed.  (found by Justin
	Sheffield at Princeton)


Special case in Penman equation

	Files affected:
	penman.c

	Description:
	Changed
		if (vpd > 0.0 && evap < 0.0)
	to
		if (vpd >= 0.0 && evap < 0.0)
	to correctly handle evap when vpd == 0.0. (found by Justin Sheffield
	at Princeton)


Rint() function not supported on all platforms

	Files affected:
	compute_dz.c, initialize_atmos.c, read_soilparam.c,
	read_soilparam_arc.c

	Description:
	Replaced rint(something) with (float)(int)(something + 0.5) to
	handle rounding without resorting to rint(), which isn't supported
	on all platforms.  (found by Justin Sheffield at Princeton)


Global parameter initialization

	Files affected:
	initialize_global.c

	Description:
	Initialize ARC_SOIL, COMPRESS, and ARNO_PARAMS to FALSE.  Also changed
	limit on loop over forcing types from hard-coded 17 to variable
	N_FORCING_TYPES.  (found by Justin Sheffield at Princeton)


Bottom soil node thickness initialization

	Files affected:
	initialize_model_state.c

	Description:
	Initialize soil_con->dz_node[Nnodes] to 0.0, since it is accessed in
	set_node_parameters().  (found by Justin Sheffield at Princeton)


Initialization of debug parameters

	Files affected:
	open_debug.c

	Description:
	Initialize debug_store_moist array when debug.PRT_MOIST is true (as
	well as under the other previously-defined conditions). (found by
	Justin Sheffield at Princeton)


Forcing data validation

	Files affected:
	read_atmos_data.c

        Description:
	Replaced NF with global_param.dt in condition checking whether forcing
	file contains enough records to cover the time range of the simulation.
	(port from 4.1.0 beta)


Model state validation

	Files affected:
	read_initial_model_state.c

	Description:
	Added check to verify that the sum of the defined nodes equals the damping depth. (port from 4.1.0 beta)