Land Processes in a High Resolution Community Climate

Land Processes in a High Resolution Community Climate Model with

Sub-Grid Scale Parameterizations

Final Progress Report Project Number DE- FG03-98ER62606

May 1, 1998 - April 30,2002

R. E. Dickinson, PI School of Earth and Atmospheric Sciences

Georgia Institute of Technology Atlanta, Georgia 30332

Report Prepared by: Andrea N. Hahmann

Institute of Atmospheric Physics The University of Arizona Tucson, Arizona 85721

July 21, 2002

Contents

1 Introduction 3

1.1 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Project milestones . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

2 Research Overview 5

2.1 The fine-mesh model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.1.1 Interaction scheme . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.1.2 Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 Topographical aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.3 Precipitation disaggregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2.4 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

A Submitted manuscripts 16

A.l Representing spatial sub-grid scale precipitation variability in a GCM . . . . . . . 16

B Publications. Abstracts. and Presentations 17

B . 1 Peer-Reviewed Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

B.2 Presentations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2

CHAPTER 1

Introduction

1 . I Objectives

The characteristics of land important for climate are very heterogeneous, as are the key atmo-

spheric inputs to land, i.e. precipitation and radiation. To adequately represent this heterogeneity,

state-of-the-art climate models should represent atmospheric inputs to land, land properties, and

the dynamical changes of land at the highest resolution accessible by climate models. The re-

search funded under this project focused on the development of an alternative approach to this

problem in which a sub-mesh is imposed on each atmospheric model grid square. This allows rep-

resentation of the land climate dynamics at a higher resolution than that achievable in the global

atmospheric models. The high spatial detail of the fine-mesh treatment provides not only a more

accurate representation of land processes to the atmospheric model, but also the opportunity for

direct downscaling of the surface climate.

The proposed project continued the development and refinement of a high-resolution land sur-

face model that is compatible for inclusion into the National Center for Atmospheric Research

(NCAR) Community Climate Model (CCM), a state-of-the-art atmospheric general circulation

model (GCM) that is used for climate simulation and prediction.

DOE Final Report July 2002

1.2 Project milestones

1. Development of the fine-mesh model interface (Fall 1998);

2. Implementation of coupling interface into CCM3.2 (Fall 1998);

3. Upgrade of coupling interface for new computer architecture - CRAY to IBM SP migration,

FORTRAN77 to FORTRAN90 migration (Spring 1999);

4. Upgrade fine-mesh input/output procedures for compatibility with CCM3.6 (Spring 1999);

5. Development of sub-grid scale parameterizations for use within the fine-mesh model (Spring

2000):

0 Development of a physical/statistical scheme to perform the spatial disaggregation of

GCM precipitation into the fine-mesh model grid;

0 Computation of incident radiation on a sloped surface according to its topographic

slope and azimuth;

6. Development of land surface datasets to be used within the fine-mesh model (Spring 2001):

0 HR vegetation and soil compatible with CCM grid;

0 Land and ocean fractions for coupling interface;

0 HR slope and azimuth (development of aggregation procedure);

0 HR elevation dataset for future use with the topographic correction;

7. Publication of research results related to the fine-mesh model:

0 Fine-mesh model development, implementation, and impact in the simulation of cli-

mate published in Journal of Climate (Hahmann and Dickinson 2001);

0 Development and sensitivity analysis of the precipitation disaggregation scheme sub-

mitted to Journal of Hydrometeorology in April 2002 (Hahmann 2002);

4

CHAPTER 2

Research Overview

The research funded by DOE'S Climate Change Prediction Program (CCPP) described in this re-

port has focused on the development of the fine-mesh model interface and its sub-grid parameteri-

zations. The following briefly describes these developments.

2.1 The fine-mesh model

The fine-mesh model represents a land surface model interface that represents sub-grid scale land

heterogeneities by a sub-mesh imposed on each atmospheric model grid. A water and energy-

conserving scheme disaggregates atmospheric conditions to the fine-mesh model grid and aggre-

gates surface fluxes back to the atmospheric grid.

The developed interface is compatible for inclusion into the National Center for Atmospheric

Research (NCAR) Community Climate Model (CCM), and uses the Biosphere- Atmosphere Trans-

fer Scheme (BATS), which is described in Dickinson et a1 (1993), as its land surface scheme. The

current model setup is as follows:

0 Atmospheric model: NCAR Community Climate Model Version 3 (CCM3), T42 (approx-

imately 2.8' x 2.8'). Figure 2.la shows the distribution of BATS land types at this model


Figure 2.1: Distribution of BATS vegetation types (1-18) at the: (a) CCM resolution (T42) and (b) fine-mesh (0.5" resolution.

resolution.

0 Land surface model: Biosphere-Atmosphere Transfer Scheme (BATS) at 0.5" x 0.5". Fig-

ure 2.lb shows the distribution of BATS land types at this model resolution.

The fine-mesh model is a high-resolution, land global model that explicitly represents the spa-

tial heterogeneities in the land surface. Fine-mesh elements with relatively uniform areas and

land cover as the only heterogeneity treated are easily implemented, are relatively efficient com-

putationally, and provide similar results to the more conventional mosaic approach. However, the

fine-mesh representation generalizes much more easily to inclusion of precipitation, solar radia-

tion, topography, soil and hydrological properties, among other factors.

6


Figure 2.2: Schematic of the relationship between fine-mesh (HR) and atmospheric model (CCM) grids. Squares labeled “0” and “E’ represent HR ocean and land points, respectively, within the CCM grid delimited by heavy black lines.

2.1 .I Interaction scheme

The atmospheric model grid (hereafter called “CCM grid”) is coupled to a fine-mesh model grid

(hereafter called “HR grid”) in two stages. First, for disaggregation, the CCM near-surface state

and downward radiative fluxes are “supplied” to the HR grid, and for aggregation, the HR grid

surface fluxes are combined and returned back to the atmospheric model. Energy and water must

exactly be conserved in both stages. Even small errors in the energy and water budget between

the surface and the atmosphere can lead to inconsistencies in the model simulation and possibly to

computational instability.

The aggregation procedure is illustrated in Figure 2.2. Surface fluxes are aggregated to a CCM

grid by area weighting the surface fluxes for all HR land grid boxes, labeled “E’, whose centers are

contained within that CCM grid box. CCM grid fluxes for boxes located along the coastline are a

linear combination of the land aggregated fluxes and the ocean fluxes. The ocean flux is computed

using the model standard ocean and sea-ice parameterizations appropriate to the CCM grid box.

2.1.2 Performance

Table 2.1 summarizes globally averaged properties for three model simulations (CCM3-LSM,

CCM3-BATS and CCM3-HRBATS). Both the annual global energy and water budgets for the

7


Table 2.1: Global and global land annual average properties. Positive fluxes correspond to those downward to the surface.

Observed CCM3 CCM3 CCM3 LSM BATS HRBATS

Global: Outgoing longwave radiation (W mP2) TOA absorbed solar radiation (W mP2) Latent heat flux (W mP2) Sensible heat flux (W mP2) Net surface solar radiation (W mP2) Net surface longwave radiation (W mP2) Precipitation (mm day-')

Global Land (149 x lo6 km2): Latent heat flux (W mP2) Sensible heat flux (W mP2) Net surface solar radiation (W mP2) Net surface longwave radiation (W mP2) Surface net radiation (W mP2) Precipitation (mm day-') Surface runoff (mm day-') 2-meter temperature (K) Surface skin temperature (K) Surface albedo (%) Surface albedo (6O0S-60"N) (%)

-234.8 238.1 -78.0 -24.0 168.0

2.69 -66.0

-

-

-

-

-

2.24 0.73

282.9 -

-

-

-236.4 236.6 -89.4 -20.5 170.9

3.07

-43.9 -31.0 144.3

76.0 2.26 0.92

280.1 280.6

24.5 20.4

-60.8

-68.3

-236.6 236.2 -90.1 -20.9 170.3

3.09

-45.3 -32.0 142.7

-65.3 77.4 2.33 0.75

280.5 280.2

25.0 20.8

-59.8

-236.1 235.7 -91.1 -20.1 169.6

-59.0 3.13

-51.8 -29.7 141.1

78.2 2.42 0.70

280.5 280.3

24.1 19.6

-62.9

three model simulations are almost indistinguishable, and are closer to each other than either is to

the observations.

Global land statistics, however, do show a few notable differences. Land surface temperatures

are warmer (by 0.4"C) in CCM3-BATS than in CCM3-LSM, consistent with larger net surface ra-

diation (by 1.4W mP2). Latent and sensible heat fluxes are larger in CCM3-BATS than in CCM3-

LSM, and precipitation in CCM3-BATS is greater by 0.07mm day-', consistent with a more vig-

orous hydrological cycle. CCM3-HRBATS shows increases in fluxes of latent heat (-6.SW mP2)

and decreases in fluxes of sensible heat (2.3W mP2) from the CCM3-BATS simulation, which are

associated to increased soil water content. These changes are accompanied by increases in land

precipitation (from 2.33 to 2.42 mm day-') and low cloud fraction (1.6%), implying in a decrease

in the net downward shortwave flux (- 1.6W mP2) and an increase in the net upward longwave

flux (-2.4W mP2) at the surface.

How surface temperature and precipitation differ geographically between the CCM3-BATS

and CCM3-HRBATS simulations is shown in Figure 2.3 and Figure 2.4 for the periods December-

8


CGM3-HRBATS - CCM3-BATS z m tsmpeietatuta

Figure 2.3: (a) DJF and (b) JJA differences (CCM3-HRBATS minus CCM3-BATS), in 2- meter temperature (“C) between model simulations. Contour interval on color bar. The shaded regions show the 95 % level significance according to the Student’s t-test of these differences

Figure 2.4: (a) DJF and (b) JJA differences (CCM3-HRBATS minus CCM3-BATS), in total precipitation (mm day-’) between model simulations. Contour interval on color bar. The shaded regions show the 95% level significance according to the Student’s t-test of these differences.

February (DJF) and June-August (JJA). These figures also show the Student’s t-test of these dif-

ferences at the 95% level. Sub-grid disaggregation parameterizations

The next step in the development of the fine-mesh model is to introduce physically based

parameterizations that allow for non-uniform distribution of the atmospheric forcing. Two such

parameterizations are presented here: a parameterization for the horizontal distribution of rainfall

and the inclusion of the geometrical properties (slope and azimuth) of the terrain.

2.2 Topographical aspects

The surface of the Earth is composed of nonuniform terrain which partially controls the amount

of incident radiation available at the surface and in turn controls regional vegetation, hydrology

and ecology. Accurate land surface-atmosphere transfer interactions may require the terrain to

9


Figure 2.5: (top) Distribution of slopes and azimuths at lkm resolution for the western United States. Figure 2.6: (bottom) Distribution of aggregated slopes and azimuths at 0.5” resolution for the western United States. The aggregation is optimized to preserve incident solar radiation.

be viewed as homogeneous sub-regions of slope and aspect, particularly over complex terrain.

Processes associated with net solar radiation and hydrology are sensitive to the degree to which the

surface slope and aspect are approximated.

The effect of topographic differences are taken into account by including the slope and the

aspect of each grid point in the computation of the incident solar radiation to be used by the land

surface scheme.

Aggregated slopes and aspects derived from the HydrolK Elevation Derivative Database, are

calculated as “radiative mean” slopes and “radiative mean” aspects, which give the equivalent

amount of radiation for each of the fine-mesh grid considered.

10


Figure 2.7: Incident solar radiation (W mP2) for a flat (pink dash line) and a surface oriented North (aspect=341") with a slope of 3.76" (green solid line). line).

Figure 2.8: Incident solar radiation (W mP2) for a flat (pink dash line) and a surface oriented South (aspect=175") with a slope of 3.56" (green solid

The original and aggregated slope and aspect data sets for the southwestern U.S. are shown in

Figure 2.5 and Figure 2.6, respectively.

The simulated direct incident solar radiation differences between inclined and flat surfaces

confirm what it is expected from theory: a north looking slope receives less radiation than a flat

surface, whereas a south looking surface receives more radiation. The diurnal variation of incident

solar radiation for a north oriented surface is shown in Figure 2.7, while the diurnal incident solar

radiation for a south oriented surface is shown in Figure 2.8.

Simulations using CCM3 coupled with the 0.5" fine-mesh model are evaluated in terms of their

snow depth and cover. A three-month ("melting season", i.e. JFM) simulation for the difference in

snow equivalent water depth between a run with implemented topography and a control run with

no topography, indicates substantial differences for the world's most important topographic areas.

To better understand the influence of topography, one needs to examine a particular GCM grid

squares where the dominant distribution of the slopes is known.

For the south oriented surface (Figure 2.9), the model shows consistent links between all quan-

tities: more solar radiation correlated with less snow, more cloud coverage correlated with more

precipitation and its associated increase in snow cover.

11


Figure 2.9: Incident solar radiation (top left), snow fraction (top right), total cloud cover (middle left), total precipitation (middle right), snow cover water equivalent depth (bottom left) and first model level air temperature (bottom right) for a grid square with dominant south orientation. Lines represent: pink (HRBATS control) and green (HRBATS + slope).

The highest snow water equivalent difference is seen where radiation-snow depth correlation is

positive, (even though the topography run receives less radiation than control one, as we expected),

the difference in cloud coverage and in temperature are consistent; there is high cloud coverage

and negative temperatures, which together, do not allow snow melting, whereas less clouds but

positive temperatures make all the radiation received (even small amounts of it) to melt the snow

accumulated.

Two important conclusions we be can draw from this preliminary study:

1. Introducing topography into the fine-mesh model coupled to CCM3 produces changes con-

sistent with the physical quantities they modify. That is most changes in parameters such as

snow cover and depth are seen over mountainous terrain.

2. In many regions multiple feedback mechanisms complicate the interpretation of the model

results.

12


Figure 2.10: Schematic representation of the CCM (green lines) and Fine-Mesh (thin blue lines) grids. Blue boxes represent grid points where precipitation occurs.

2.3 Precipitation disaggregation

In addition to the inherent sub-grid nature of surface characteristics, the inputs, or “forcings”,

from the atmosphere to the land surface are also highly heterogeneous. The most variable of these

forcings is the precipitation field.

One approach in incorporating sub-grid scale precipitation variability in land surface models

is to consider its statistical representation within the GCM grid square. Most models that use

this approach assume that rainfall is patchy, and falling only on a small fraction of the GCM grid

square. The statistical distribution of rainfall is then incorporated into the equations controlling

hydrological processes such as the runoff generation and evapotranspiration.

Within the fine-mesh model we have chosen to explicitly represent rainfall variability within a

GCM grid square. We assume that the majority of precipitation events occur over a small fraction

of the entire grid square. For simplicity, we also assume that the spatial distribution of precipitation

is uncorrelated in time.

The procedure used within the fine-mesh model follows. Consider a GCM grid square com-

posed of N x M sub-grid squares. The disaggregation scheme first chooses a single grid random

point (2 , j ) among all possible sub-grid squares. The amount of rainfall received by this point is

determined by an exponential distribution of rainfall rates within the GCM grid square. This pro-

13


1

Figure 2.11: CCM3/HRBATS + disaggregation scheme (p = 0.2) minus control differences in the first model layer temperature (K) for DJF (top) and JJA (bottom). Contour interval of 1.O”C. The shaded regions show the 95% level significance according to the Student’s t-test of these differences.

Figure 2.12: CCM3/HRBATS + disaggregation scheme (p = 0.2) minus control differences precipitation (mmday-l) for DJF (top) and JJA (bottom). Contour interval of 0.5, 1.0, and 2.0mm day-’. The shaded regions show the 95% level significance according to the Student’s t-test of these differences.

cess is iterated until a fraction p of the GCM grid square (or, p x N x N sub-mesh grid points) is

covered with rainfall. Fine-mesh precipitation amounts are then normalized to conserve the GCM-

produced precipitation. A schematic representation of the scheme is presented in Figure 2.10.

The exponential distribution of fine-mesh rainfall is highly attractive because it depends on a

single parameter, the “wetting” area, which could potentially be estimated from observations of

rainfall or the thermodynamical structure of the large-scale environment.

The differences in surface temperature and precipitation between the control (CCM3-HRBATS)

and CCM3-HRBATS with p=0.2 simulations are presented in Figure 2.1 1 and Figure 2.12. Changes

in near-surface temperatures include cooler land (1°C) over most of the mid and high-latitude con-

tinents in the Northern Hemisphere during D F . In the tropics, the reductions in evapotranspiration

lead to significant warming (1-2°C) in both DJF and JJA.

14


Changes in precipitation are mainly confined to the tropical area. During DJF, there are large

and significant differences in precipitation over tropical land, with decreases of the order of 1-2

mm day-' over the Amazon and tropical Africa. During JJA, the largest differences in precipita-

tion migrate north, with increases over northern South America and Central America and India,

and an area of significant decreases in precipitation over the central United States.

Results from this aspect of our research have recently been submitted for publication. A copy

of the abstract of this manuscript (Hahmann 2002) is available in Appendix A.

2.4 References

Hahmann, A. N., 2002: Representing spatial sub-grid scale precipitation variability in a GCM. J.

Hydrometeorology. Submitted.

Hahmann, A. N., and R. E. Dickinson, 2001: A fine-mesh land approach for general circulation

models and its impact on regional climate. J. Climate, 14, 1634-1646.

1s

APPENDIX A

Submitted manuscripts

A.l Representing spatial sub-grid scale precipitation variability in a GCM

Manuscript submitted to Journal of Hydrometeorology

Abstract

A method for spatial disaggregation of rainfall to be used within a general circulation model

(GCM) is presented. The method is based on the assumption of exponential distribution of rain-

fall rates that are a function of the area of the GCM grid square where precipitation is occurring.

The climate simulated using the precipitation disaggregation scheme shows considerable changes

from the simulated control climate with overall warmer and drier tropical land and warmer high

latitude continental areas. Changes in the simulated climate can be explained by direct local land-

atmosphere interaction processes and remote responses to changes in tropical heating linked to dif-

ferences in precipitation. In the tropics, the relationship between decreases in rain-covered fraction

and decreases in GCM surface evaporation and precipitation follows an exponential relationship.

16

APPENDIX B

Publications, Abstracts, and Presentations

B.l Peer-Reviewed Publications

Buermann, W., J. Dong, X. Zeng, R. B. Myneni, and R. E. Dickinson, 2001: Evaluation of the

utility of satellite-based vegetation leaf area index data for climate simulations. J. Climate, 14,

3536-3550.

Chen, M., X. Zeng, and R. E. Dickinson, 1998: Adjustment of GCM precipitation intensity over

the United States. J. Appl. Meteor., 37, 876-887.

Dickinson, R. E. and K. J. Schaudt, 1998: Analysis of timescales of response of a simple climate

model. J . Climate, 11, 97-106.

Dickinson, R. E., M. Shaikh, R. Bryant, and L. Graumlich, 1998: Interactive canopies for a climate

model. J . Climate, 11,2823-2836.

Dickinson, R. E., J. A. Berry, G. B. Bonan, G. J. Collatz, C. B. Field, I. Y. Fung, M. Goulden, W.

A. Hoffman, R. B. Jackson, R. Myneni, P. J. Sellers, and M. Shaikh, 2001: Nitrogen controls

on climate model evapotranspiration. J. Climate, 15,278-295.

17


Hahmann, A. N., 2002: Representing spatial sub-grid scale precipitation variability in a GCM. J.

Hydrometeorology. Submitted.

Hahmann, A. N., and R. E. Dickinson, 2001: A fine-mesh land approach for general circulation

models and its impact on regional climate. J. Climate, 14, 1634-1646.

Koster, D. K., P. A. Dirmeyer, A. N. Hahmann, R. Ijpelaar, L. Tyahla, P. Cox, , and M. J. Suarez,

2002: Comparing the degree of land-atmospheric interaction in four atmospheric general cir-

culation models. J. Hydrometeorology, 3, 363-375.

Martin M., Dickinson R. E, Yang Z. L., 1999: Use of a coupled land surface general circulation

model to examine the impacts of doubled stomatal resistance on the water resources of the

American southwest. J. Climate, 12, 3359-3375.

Morrill, J. C., R. E. Dickinson, and A. N. Hahmann, 1999: Sensitivity of a land surface model

to the diurnal distribution of downward longwave radiation. J. of the Meteorological SOC. of

Japan, 77,265-279.

Polcher, J., B. McAvaney, P. Viterbo, M.-A. Gaertner, A. N. Hahmann, J.-E Mahfouf, J. Noilhan,

T. Phillips, A. Pitman, C. A. Schlosser, J.-P. Schultz, B. Timbal, D. Verseghy, and Y. Xue,

1998: A proposal for a general interface between land-surface schemes and general circulation

models. Global and Planetary Change, 19, 263-278.

Seth A., R. C. Bales, and R. E. Dickinson, 1999: A framework for the study of seasonal snow

hydrology and its interannual variability in the alpine regions of the Southwest. J. Geophys.

Res, 104,No. D18,22,117-22,135.

Wei, X, A. N. Hahmann, R. E. Dickinson, Z.-L. Yang, X. Zeng, K. J. Schaudt, C. B. Schaaf, N.

Strugnell, 2001: Comparison of albedos computed by land surface models and their evaluation

against remotely sensed data. J. Geophys. Res., 106, No. D18,20,687-20,702.

Yang, Z.-L., R. E. Dickinson, A. N. Hahmann, M. Shaikh, G.-Y. Niu, X. Gao, R. C. Bales, S.

Sorooshian and J. M. Jin, 1999: Simulation of snow mass and extent in global climate models.

Hydrological Processes, 13,2097-21 13.

18


Zeng, X. and R. E. Dickinson, 1998: Effect of surface sublayer on surface skin temperature and

fluxes. J. Climate, 11,537-550.

Zeng, X., Y.-J. Dai, R. E. Dickinson, and M. Shaikh, 1998: The role of root distribution for land

climate simulation. Geophys. Res. Lett., 25,4533-4536.

Zeng, X., R. E. Dickinson, A. Walker, M. Shaikh, R. S. DeFries, and J. Qi, 2000: Derivation and

evaluation of global 1-km fractional vegetation cover data for land modeling. J. Appl. Meteor.,

39,826-839.

Zeng, X., 2001: Global vegetation root distribution for land modeling. J. Hydrometeor., 2, 525-

530.

Zeng, X., M. Shaikh, Y. Dai, R. E. Dickinson, and R. Myneni, 2002: Coupling of the Common

Land Model to the NCAR Community Climate Model. J . Climate, 14, 1832-1 854.

B.2 Presentations

Land-Ocean-Atmosphere Interactions Over South America. Presented by A. Hahmann at the Sixth

International Conference on Southern Hemisphere Meteorology and Oceanography, Santiago,

Chile, April 2000.

Tropical Deforestation and Climate: A Testbed for Understanding Land Surface Processes. Pre-

sented by A. Hahmann as an Invited talk to the GEWEX/INSU Workshop on “Modeling

land-surface-atmosphere interactions and climate variability”, Gif-sur-Yvette, France, October

1999.

High-Resolution Simulations of the Global Climate - Validation of the Monsoon System over

the Southwest United States and Northern Mexico. Presented by A. Hahmann at the 22nd

General Assembly of the International Union of Geodesy and Geophysics, Birmingham, United

Kingdom, July 1999.

19


The Diurnal Cycle of Precipitation in the Amazon Basin: Sensitivity Experiments with a Global

Fine-Mesh Land Model. Presented by A. Hahmann at 23rd Conference on Hurricanes and

Tropical Meteorology. 1999 AMS Annual Meeting, Dallas, Texas, January 1999.

High-Resolution Simulations of the Global Climate: Validation of the Hydrological Cycle of the

Amazon and Southwest U.S. Presented by A. Hahmann at the 10th Symposium on Global

Change Studies, 1999 AMS Annual Meeting, Dallas, Texas, January 1999.

Land Processes in a High Resolution Community Climate Model with Sub-Grid Scale Parame-

terizations. Presented by A. Hahmann at the Climate Change Prediction Meeting, Phoenix,

Arizona, October, 1998.

20

Documents

Land Processes in a High Resolution Community Climate