Estimate of acid deposition through fog using numerical models in - - PowerPoint PPT Presentation

Estimate of acid deposition through fog using numerical models in the Kinki Region of Japan

Hikari Shimadera1, 2, Akira Kondo1, Akikazu Kaga1, Kundan Lal Shrestha1, 2, Yoshio Inoue1

1 Graduate School of Engineering, Osaka University 2 JSPS Research Fellow

H13-92

1

Introduction Background

Acid deposition has been widely recognized as a regional environmental problem, and has caused damage to sensitive ecosystems Fog deposition can lead to considerable amount of acid deposition in mountainous forest areas

– Ionic concentrations in fog are much higher than those in rain – Fog water deposition through interception by vegetation can be an important part of the hydrologic budget of forests

Few fog monitoring sites exist and fog is highly variable according to region

2

Introduction Objective

Establish a method to estimate spatial distribution of the amount of acid deposition including fog deposition

Approach

2-dimensional model to predict fog water deposition (FDM) was developed and verified FDM was applied with meteorology and air quality model to estimate acid deposition in Kinki Region of Japan

2-dimensional Fog Deposition Model (FDM) Description of FDM Features of FDM Comparison of FDM with field measurement

3

Description of FDM

u u z a C z u K t t u

S d M

1 , for ) ( ˆ Z h z Z Z a h SAI z a

fc fc S

, 2 1 2 1 exp 1 1 ) ( ˆ

2 2

Z Z Z Z a Z a

m m m

1 ) ( ˆ

1

dZ Z a

KM: eddy diffusivity, Cd: drag coefficient, aS: surface area density

4

Equation of mean motion

1 for 1 for 2 4 1 1

2 m

Z SAI: surface area index (= LAI+NLAI), hfc: height of forest canopy λ: parameter by Kondo and Arakashi (1976, Boundary-Layer Meteorol., 10 (3), 255)

Description of FDM

Dep z LWC K z z LWC w x LWC u t LWC

M

LWC u z a f Dep

IM L L

, St St

IM

fL: portion of the effective leaf area, aL: leaf area density, εIM : impaction efficiency of fog droplet

L A p W

d u d St 9

2

α, β, γ: 5.0, 1.05 and 1 for needle leaf 0.5, 1.90 and 5 for broad leaf dL: characteristic leaf length (=0.001 m for needle leaf and 0.030 m for broad leaf) dp: mean diameter of fog droplet (= 17.03LWC×10-3+9.72×10-6 m) by Katata et al. (2008, J Appl. Meteorol. Climatol., 47 (8), 2129)

5

Equation of liquid water content of fog (LWC)

z= 0 ubc FRF 1-FRF forest area non-forest area z= Δz×nz z = hfc x = 0 x = Δx×nx LWCbc

SAI/hfc aS(z) u(z)

Description of FDM

6 λ = 1 λ = 2 λ = 3 λ = 4 λ = 5

– Forests are allocated to the computational area from its horizontal edges according to fraction of forest (FRF) – Vertical distributions of aS(z) vary with λ, and u(z) vary with aS(z) – FDM predicts steady state fog deposition velocity (VDep = fog water deposition flux/LWCbc) for each run – FDM configuration λ = 3, NLAI = 0.5, Δx = 40 m, nx = 50, Δz = 1.5 m, nz = 30

10 1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 2 4 6 8 10

Features of FDM

0.1 0.3 1 3 10

Mean VDep (m s-1) ubc (m s-1)

7

Sensitivity to parameters

– Since εIM increases with u, VDep increases with ubc – When forest areas are thin, VDep considerably increases with LAI – When forest areas are dense, VDep does not very increase or can decrease with an increase in LAI because of large drag force – FDM configuration LWCbc = 0.0003 kg m-3, hfc = 18 m, FRF = 0.96, needle-leaved forest

LAI

0.0 0.2 0.4 0.6 0.8 1.0 1.2

400 800 1200 1600 2000 2000

Features of FDM

Horizontal distribution of VDep

8

FRF x (m)

– VDep at the windward edge of forest is the largest in every cases – Ratio of (Vdep at edge/VDep in inner forest) increases with increasing width of the gap between forest areas

ubc x = 0 x = 2000 m

largest VDep

– FDM configuration ubc = 10 m s-1, LWCbc = 0.0003 kg m-3, hfc = 18 m, FRF = 0.96, LAI = 3, needle-leaved forest

VDep (m s-1)

0.24 0.48 0.96

Measurement site

– Burkard et al. (2003, Atmos. Environ., 37 (21), 2979) measured the turbulent fog water flux by the eddy covariance method at 45 m on a tower (15 m above the forest canopy) – The measurement site is situated at 690 m on the south slope of the Lägeren Mountain, ~15 km northwest of Zurich, Switzerland – The vegetation cover around the site is mixed forest dominated by beech and Norway spruce – FDM configuration hfc = 30 m, LAI derived from dataset of MODIS LAI product, FRF = 1, 50 % of needle-leaved and 50 % of broad-leaved trees

Comparison with field measurement

9 Eddy covariance measurement Lägeren forest (cited from http://www.gl.ethz.ch/infrastructure/research_sites/switzerland/laegeren)

2 4 6 8

FDM Measurement

1.0E-04 1.0E-03 1.0E-02 1.0E-01 1.0E+00 2 4 6 8 10

Fog water flux and Vdep

Comparison with field measurement

Accumlated turbulent fog water flux (mm) FDM Measurement

• Sep. Oct. Nov. Dec. Jan. Feb. Mar.

2001 2002 VDep (m s-1) ubc (m s-1)

10

– Total fog water flux in FDM (7.3 mm) agreed with that in the measurement (7.4 mm) – As VDep in FDM strongly depend on u, FDM underestimated fog water flux when measured VDep was large despite low ubc, and overestimated when measured VDep was small despite high ubc

Estimate of fog deposition in Kinki Region of Japan Modeling system Modeling domain Air quality prediction in March 2005 Forest data Fog water deposition and corresponding NOY deposition in March 2005

11

* NOY = NO + NO2 + NO3 + N2O5 + HNO3 + HONO + aerosol nitrate

Modeling system

Japan: EAGrid2000-JAPAN Other Asian countries Anthropogenic: 2006 INTEX-B Asia emission NH3: REAS 2005 prediction Biogenic VOCs: EAGrid2000 Biomass Burning: Streets et al. (2003) Volcanic SO2: Observed value at Miyakejima Andres and Kasgnoc (1998)

Meteorology Field Emission Data Atmospheric concentration Dry/Wet deposition

MM5 v3.7

(The 5th-Generation NCAR/ Penn State Mesoscale Model)

Fog deposition

CMAQ v4.7

(The U.S. EPA’s Community Multiscale Air Quality model)

FDM

Ionic concentration in fog water Emission Processor

CCTM

(CMAQ Chemical Transport Model)

MCIP v3.4

(Meteorology-Chemistry Interface Processor) CMAQ default concentration profiles Photolysis rate table NCEP.FNL (D1) GPV-MSM (D2) Forest data

MM5/CMAQ modeling domain

13

100 200 300 400 500 600 700 850 1000 1200 1500 2000

Elevation [m]

Horizontal resolution

– D1: 54km-grid cells (105×81) – D2: 18km-grid cells (72×72) – D3: 6km-grid cells (99×99) – D4: 2km-grid cells (126×126)

Vertical resolution

– 24 layers from surface to 100hPa (the middle height of the 1st and 2nd layers are ~15 and 50m)

• Mt. Rokko

D4 D1 D2 D3

for estimate of fog deposition

14

Air quality prediction in March 2005

0.3 0.5 0.7 1.0 1.3 1.7 2.2 2.8 3.5 4.3 5.2 7.0 10.0 15.0 5 7 10 13 17 22 28 35 43 52 70 100 150 250

Aerosol NO3

• Spatial distributions of monthly mean NO3
• concentrations

NO3

• in rain

NO3

• in fog

[μmol L-1]

5 7 10 13 17 22 28 35 43 52 70 100 150 250

[μmol L-1] [μg m-3]

15

Air quality prediction in March 2005

Aerosol NO3

• Observation [μg m-3]

Prediction [μg m-3]

20 40 60 80 20 40 60 80

Comparisons of predictions with observations

NO3

• in rain

2 4 6 2 4 6

Observation [μmol L-1] Prediction [μmol L-1]

NO3

• in fog at Mt. Rokko

1.E+0 1.E+1 1.E+2 1.E+3 1.E+4 1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31

NO3

• in fog [μmol L-1]

Date of March 2005 Observation

Prediction

Forest data

FRF, forest class and LAI were obtained from dataset of MLIT in Japan, MOE in Japan and MODIS LAI product Forest areas account for 65 % of the land areas and 95 % of the mountainous areas with elevation > 500 m Needle-leaved (DN + EN) forest account for 67 % of the forest area and broad-leaved (DB + EB) forest account 33 % Because March is before or at the beginning of the vegetation growing season, most of the forest areas tend to be thin North-eastern areas covered with deciduous broad-leaved (DB) forest show the lowest LAI

16

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.3 0.6 0.9 1.2 1.5 1.8 2.1 2.4 2.7 3.0

FRF

DN EN DB EB

LAI in March 2005 Forest class

[m2 m-2]

Fog water deposition in March 2005

17

10 40 80 120 160 220 280 360 450 1 2 3 4 5 6 8 10 12 15 1 4 8 12 16 22 28 36 45

Fog frequency Fog water deposition Rainfall

[mm] [mm] [%]

Fog frequency, fog water deposition and rainfall generally increased with increasing elevation While fog frequency and rainfall were the highest in the north- eastern area dominantly covered with deciduous broad-leaved forest, fog water deposition was not due to the thin vegetation cover Ratios of (Fog water deposition/Rainfall) reached up to 23 % (mean = 3 %) in mountainous areas The ratio may change in the vegetation growing season

NOY deposition through fog in March 2005

18

Fog deposition Rainfall

[mmol m-2] [mmol m-2]

Ratios of (NOY depositions through fog/NOY depositions through rain) reached up to 97 % (mean = 8 %) in mountainous areas Contribution of fog deposition to NOY depositions was larger than that

• f dry deposition in some mountainous areas

Fog water deposition is an important pathway for acid deposition in some mountainous areas in Kinki Region, Japan NOY deposition corresponding to

0.6 1.2 2.0 3.2 4.8 7.0 10 15 22 32 0.06 0.12 0.20 0.32 0.48 0.70 1.0 1.5 2.2 3.2

[mmol m-2]

0.6 1.2 2.0 3.2 4.8 7.0 10 15 22 32

Dry deposition

19

Conclusions

Fog deposition model (FDM) was developed and verified

– VDep calculated by FDM considerably varied with u and parameters on forest – Despite some discrepancies between FDM and the measurement, FDM captured the total fog water deposition

Fog deposition was estimated with FDM and MM5/CMAQ in Kinki Region, Japan in March 2005

– Fog water deposition can contribute significantly to acid deposition in some mountainous areas – Long-term prediction (1year ~) is required for further study, because contribution of fog deposition vary with seasonal variations in meteorology, air quality and vegetation structure

Thank you