Mass Deposition Fluxes of Asian Dust to the Bohai Sea and Yellow Sea from Geostationary Satellite MTSAT: A Case Study

Abstract

Windblown dust aerosol plays an important role in marine ecosystems once they are deposited and dissolved. At present, methods for estimating the deposition flux are mainly limited to direct measurements or model outputs. Additionally, satellite remote sensing was often used to estimate the integral dust column concentration (DCC). In this paper, an algorithm is developed to estimate the mass deposition fluxes of Asian dust by satellite. The dust aerosol is identified firstly and then the DCC is derived based on the relationships between the pre-calculated lookup table (LUT) and observations from Japanese geostationary Multi-functional Transport Satellites (MTSAT). The LUT is built on the dust cloud and surface parameters by a radiation transfer model Streamer. The average change rate of deposition is derived, which shows an exponential decay dependence on transport time along the pathway. Thus, the deposition flux is acquired via integrating the hourly deposition. This simple algorithm is applied to a dust storm that occurred in the Bohai Sea and Yellow Sea from 1 to 3 March 2008. Results indicate that the properties of the dust cloud over the study area changed rapidly and the mass deposition flux is estimated to be 2.59 Mt.

1. Introduction

Dust aerosol carried by dust storm events from Asia not only damage to the local ecosystems and human society, but also have large impacts on downwind areas such as Japan, Korea, and even the West Pacific Ocean through long-range transport [1]. Dust aerosol in the atmosphere play an important role in the radiative forcing of the atmosphere via scattering and absorbing shortwave and longwave radiation [2]. Once deposited and dissolved in the ocean, it modifies biogeochemistry and carbon export [3]. Especially in the high nutrient low chlorophyll (HNLC) open ocean regions, the availability of iron is the limiting factor for phytoplankton growth [4]. It is reported that direct bio-available soluble iron occupies 0.01%–80% of the dust aerosol [5]. The deposition fluxes of particulate Fe and P during the dust storm events associated with precipitation were about 500–1000 times of the daily averaged flux during non-dust days [6]. Tan and Wang [7] found that the iron deposited by the severe dust episode could have increased the chlorophyll a concentration in the southern Yellow Sea by 10%–68%.Yuan and Zhang [8] also showed a strong correlation between the dust events and biological productivity through the oceanic sediment-trap substances. However, significant uncertainties still remain in understanding the whole dust process, including generation, long-range transport, deposition, and their impact on ocean.

In this paper, we focus on the estimation of dust mass deposition fluxes over the ocean. The dust aerosol sinks into the ocean through wet and dry deposition in the transport passage. The deposition in the ocean plays an important role in the mineral dust cycle. The development of the dust deposition flux estimation methods was summarized, and they were compared to each other by Niedermeier et al. [9]. The dust deposition flux is mainly from direct measurements or model outputs. Each method has its own advantages and disadvantages. The direct measurement method is easy to calculate. However, the samplings are sensitive to wind erosion and other perturbations, which need to be corrected for the efficiency of the collection sampler. Furthermore, direct measurements are sparse and non-uniform, especially in the open oceans because of the prohibitive costs as well as the episodic nature of atmospheric deposition into the ocean. Therefore, various global [10,11] and regional [12,13] dust models have been developed to estimate the emissions, transport and deposition since the late 1980s. In contrast to global models, regional models provide better results with higher spatial resolution and lower uncertainties. Unfortunately, the synchronous meteorological and hydrological fields are often very scarce, which are necessary for the models. Current model estimates of mass concentrations of dust can vary from a factor of two at the global scale to at least one order of magnitude in regional model studies [14]. Sometimes, the results from different model simulations are different from each other when performing the same test [13].

On the one hand, because of the lack of continuous and simultaneous observational data, dust deposition over the ocean is still poorly understood, which badly needs quantitative evaluation research. On the other hand, satellites can provide continuous dust monitoring over large areas now and the development of inverse data calculation methods can produce quantitative estimates of dust column concentration (called dust burden or dust load). An algorithm for the retrieval of volcanic particle size and optical thickness to derive the column concentration from the Advanced Very High Resolution Radiometer (AVHRR) thermal infrared observations was developed by Wen and Rose [15]. A similar method was applied to a 7 April 2001 dust storm over northern China using the Moderate Resolution Imaging Spectroradiometer (MODIS) data [16]. The wide spectral range of visible and near infrared technique can be used to retrieve dust load over ocean was demonstrated by Tanré et al. [17]. The first example of MODIS data to distinguish dust from smoke and maritime aerosols and to evaluate the African dust column concentration was shown by Kaufman et al. [18]. A retrieval algorithm to obtain the optical thickness, dust particle effective radius and integral dust column density simultaneous, using three MODIS thermal IR channels was developed by Zhang et al. [19]. Polar satellite observations are widely used to estimate the dust column concentration. However, it is insufficient to estimate the dust deposition by a snapshot of polar satellite observations as dust storms evolve with time and space rapidly.

In this study, a new algorithm is developed to estimate the deposition flux of a dust event using the geostationary satellite data. The organization of this paper is as follows. In Section 2, a brief overview of the Multifunctional Transport Satellites 1R (MTSAT-1R) data and Streamer model are provided. The detail of dust deposition flux estimation algorithm is presented in Section 3. Section 4 explains the technique applied to the dust event over the Bohai Sea and Yellow Sea from 1 to 3 March 2008. Conclusions are presented in Section 5.

2. Data and Radiative Transfer Models

2.1. Multifunctional Transport Satellites 1R (MTSAT-1R) Observations

The Multi-functional Transport Satellite1R (hereafter MTSAT-1R) of Japan Meteorological Agency was launched on 26 February 2005. It is located 35,800 km above the equator at 140 degrees east longitude and provides wide coverage for the Pacific Ocean. MTSAT-1R carries a five-channel imager called Japanese Advanced Meteorological Imager (JAMI), providing nominal half-hourly images. JAMI consists of one visible channel and four infrared (IR) channels. The characteristics of these channels are listed in Table 1 (for more details, see //www.jma.go.jp/jma/jma-eng/satellite/).

Table 1. The characteristics of MTSAT channels.

Channel	Wavelength /μm	Quantization Levels/bit	Space Resolution/km
Visible (VIS)	0.55–0.90	10	1
Thermal IR 1 (IR1)	10.3–11.3	10	4
Thermal IR 2 (IR2)	11.5–12.5	10	4
IR water vapor (IR3)	6.5–7.0	10	4
Mid-IR (R4)	3.5–4.0	10	4

Table 1. The characteristics of MTSAT channels.

Channel	Wavelength /μm	Quantization Levels/bit	Space Resolution/km
Visible (VIS)	0.55–0.90	10	1
Thermal IR 1 (IR1)	10.3–11.3	10	4
Thermal IR 2 (IR2)	11.5–12.5	10	4
IR water vapor (IR3)	6.5–7.0	10	4
Mid-IR (R4)	3.5–4.0	10	4

Two thermal infrared data, IR1 and IR2, are used to monitor the transport of dust storm events and quantify the depositions during the process. IR1 and IR2 of MTSAT-1R are well calibrated and show no serious systematic errors or biases with respect to the corresponding channels of MODIS [20].

2.2. Streamer

There are many radiative transfer models, such as LOW resolution TRANsmission Model (LOWTRAN), MODerate resolution atmospheric TRANsmission (MODTRAN), Second Simulation of the Satellite Signal in the Solar Spectrum (6S), and Streamer, widely used in remote sensing applications. Streamer, which is based on plane-parallel theory, possesses the ability to simulate radiances for a variety of cloud atmospheric situations [21]. It is available freely as well as being efficient, flexible and user-friendly. In this paper, Streamer is selected to simulate the satellite radiance of dust aerosol. In Streamer simulation, each computed radiance or reflectance is done for a scene by using the discrete ordinate solver, DISORT (Discrete ordinates radiative transfer) [22]. There are seven standard atmospheric profiles in the Streamer: tropical, mid-latitude summer, mid-latitude winter, subarctic summer, subarctic winter, arctic summer and arctic winter. User-defined profiles can also be used. Dust aerosol optical and geometrical thickness, or the optical properties and phase function, may be specified as input. In addition, in order to better simulate the satellite observations, the response function of MTSAT1R is integrated into the Streamer model.

3. Methodology

According to the retrieval method of Zhang et al. [23], the dust optical thickness and particle size are retrieved first. Dust column density is then derived from dust optical thickness and particle size. The dust deposition is calculated based on the hourly MTSAT column dust density (Figure 1).

Figure 1. Flow chart of the dust deposition flux estimation process.

Figure 1. Flow chart of the dust deposition flux estimation process.

The basic theory and algorithm can be briefly described as follows. Satellite sensor observations, through a semitransparent dust cloud, include two parts: radiance transmitted from the underlying surface and radiance emitted from the dust cloud. This can be expressed as the following equation [15]:

$$I_i=t_i({\text{r}}_{\text{e}},{\tau }_{\text{du}})(B(T_s)-B(T_{du}))+(1-R_i({\text{r}}_{\text{e}},{\tau }_{\text{du}}))B(T_{du})$$

(1)

where I_i is the radiance centered at wavelength λ_i, T_s is the brightness temperature of the surface, T_du is the equivalent temperature of the top of the cloud, B is the Planck function, t is the transmissivity, R is the reflectivity, r_e is the effective radius of the dust, and τ_du is the dust optical depth. r_e and τ_du are defined as follows:

$${\tau }_{\text{du}}=\int {\text{Q}}_{ext}\pi {\text{r}}^{2}n(r)dr$$

(2)

$${\text{r}}_{\text{e}}=\frac{\int {\text{πr}}^3\text{n}(\text{r})\text{dr}}{\int {\text{πr}}^2\text{n}(\text{r})\text{dr}}$$

(3)

where n(r) is the particle size distribution, and Q_ext is the light extinction efficiency decided by Mie scattering. For the large size parameter of dust, the Q_ext is oscillating near 2. t and R are functions of r_e and τ_du. That is to say, r_e and τ_du can be retrieved simultaneously by a pair of radiances (I_i, I_j). The infrared observations of MTSAT have been transformed to brightness temperatures instead of radiance, although they can be transformed easily by Plank Law. Thus, the radiative transfer model Streamer is used to simulate the theoretical brightness temperatures at various dust cloud and surface parameters with specific aerosol physical parameters. The aerosol physical parameters, such as the single scattering albedo, asymmetry factor, and extinction coefficient ratio of the dust layer, are calculated first by Mie scattering theory based on several assumptions [24]. These assumptions are: (1) dust particles are spherical; (2) dust particle size distribution is lognormal; and (3) complex refractive index is calculated based on HITRAN 2008 dataset under different relative humidity conditions (0, 30%, 50%, 70%, and 95%). To simplify the calculation, we assume that the underlying surface is homogeneous and the dust aerosol is a single layer parallel to the surface. The lookup table (LUT) is calculated with seven T_s (from 265 K to 295 K in 5 K increments), seven T_du (from 260 K to 290 K in 5 K increments), 19 optical thicknesses (from 0.3 to 1 in 0.1 increments and 1 to 15 in increments of 1), and 29 particle effective radii (from 0.1 to 2 in 0.1 increments and 2 to 10 in increments of 1). The main input parameters used to generate the LUT are summarized in Table 2. To make the simulation more realistic, user-defined atmospheric profiles, instead of the standard atmospheric profiles, are used in Streamer. The atmospheric profiles of pressure, temperature and humidity are defined based on the sounding observations (http://www.weather.uwyo.edu/upper air/sounding.html) on the transport pathway during the dust event. Emissivity type is the surface emissivity control. Generally, ocean emissivity lies between 0.96 and 0.99 in infrared bands. In this study, 0.98 is chosen for infrared bands calculation. The height of the long-range transport of Asia to the China Seas were usually in heights between 0.5 km and 1.5 km [25]. According to this, the bottom height is set to 0.5 km.

Table 2. Streamer input parameters.

Parameter	Value
Atmospheric profiles	User-defined by radiosonde data
surface type	water
emissivity	0.98
Surface temperatures (K)	(265~295 (5)),bracketed text indicates step
dust temperatures (K)	(260~290 (5)), bracketed text indicates step
dust cloud bottom height (km)	0.5
optical thickness	(0.3~1.0 (0.1), 1.0~15.0 (1))
effective radius (μm)	(0.1~2.0 (0.1), 2~10 (1))

Table 2. Streamer input parameters.

Parameter	Value
Atmospheric profiles	User-defined by radiosonde data
surface type	water
emissivity	0.98
Surface temperatures (K)	(265~295 (5)),bracketed text indicates step
dust temperatures (K)	(260~290 (5)), bracketed text indicates step
dust cloud bottom height (km)	0.5
optical thickness	(0.3~1.0 (0.1), 1.0~15.0 (1))
effective radius (μm)	(0.1~2.0 (0.1), 2~10 (1))

The MTSAT brightness temperatures are then compared with pre-calculated values from LUT until the best (least-squares) fit is obtained. The optical thickness τ_du and effective radius r_e can be retrieved form the best fit. Then, dust mass column concentration M_du is derived from the optical thickness and effective radius, using the following relationships between the dust mass and volume to the dust size distributions:

$${\text{M}}_{\text{du}}=\frac{4\text{ρ}}{3}\int \pi {\text{r}}^{3}n(r)dr$$

(4)

Combining Equations (2)–(4),

$${\text{M}}_{\text{du}}=\frac{2\text{ρ}}{3}{\text{r}}_e{\text{τ}}_{\text{du}}$$

(5)

Thus, the dust column density is widely estimated from polar sun synchronous satellites. However, it is insufficient to estimate the deposition from the column dust density only at a single time due to the high temporal and spatial variability, and the deposited rate is usually unknown. The continuous geostationary satellites observations with high temporal resolution provide the opportunity to estimate the ratio of deposition $\overline{v_{\text{du}}}$. The $\overline{v_{\text{du}}}$ is determined by derivative of an exponential model described in Section 4.3, which is obtained by means of regression from the selected mean dust column density $\overline{{\text{M}}_{\text{du}}}$ over the water. Thus, the dust deposition at time i can be estimated as follows,

$$\overline{{\text{M}}_{\text{de}}[i]}=\overline{{\text{M}}_{\text{du}}[i]}\times \overline{{\upsilon }_{\text{du}}[i]}\times S_{du}[i]$$

(6)

where $S_{du}$ represents the cover area of the dust. The total deposition amount M_de is the sum of $\overline{{\text{M}}_{\text{de}}\left[\text{i}\right]}$ throughout the event,

$${\text{M}}_{\text{de}}={\displaystyle \sum _{i=1}^T\overline{{\text{M}}_{\text{de}}[i]}}$$

(7)

where T is the duration of the dust event, defined once the dust cloud transport to the study region until they pass through entirely.

4. Study Case

In this study, the MTSAT observation of the dust storm on 1–3 March 2008 is selected as the case for examining the dust deposition flux estimation algorithm.

4.1. Dust Detection

The dust cloud over the Bohai Sea and Yellow Sea on 1–3 March 2008 was discriminated first based on the detection algorithm developed by Tu et al. [26]. In the MTSAT brightness temperature difference (BTD) images (Figure 2), the strong negative BTD signals indicate the larger amounts of dust aerosol over the sea. The strong negative BTD results from the absorption effects of mineral dust. The dust aerosol was verified by the near simultaneous RGB true color composition images results from National Oceanic and Atmospheric Association (NOAA) series (Figure 3).

Figure 2. MTSAT brightness temperature difference (BTD) images identified the evolution of the dust event on 1–2 March 2008; (a) 200803010130; (b) 200803010530; (c) 200803010730; (d) 200803020130; (e) 200803020430; (f) 200803020730.

Figure 2. MTSAT brightness temperature difference (BTD) images identified the evolution of the dust event on 1–2 March 2008; (a) 200803010130; (b) 200803010530; (c) 200803010730; (d) 200803020130; (e) 200803020430; (f) 200803020730.

Figure 3. RGB true color composition images from the NOAA series for the comparison of the dust detection results corresponding to Figure 2; (a) 200803010150(NOAA17); (b) 200803010502(NOAA18); (c) 200803010747(NOAA16); (d) 200803020127(NOAA17); (e) 200803020452(NOAA18); (f) 200803020735(NOAA16).

Figure 3. RGB true color composition images from the NOAA series for the comparison of the dust detection results corresponding to Figure 2; (a) 200803010150(NOAA17); (b) 200803010502(NOAA18); (c) 200803010747(NOAA16); (d) 200803020127(NOAA17); (e) 200803020452(NOAA18); (f) 200803020735(NOAA16).

On 1 March 2008, a dust storm originated from Mongolia, spread to China over the Inner Mongolia and Loess Plateau events. Then, the evolution of the dust is indicated by the yellow color relative to the white cloud systems on true color images of the NOAA series satellites. It began entering Bohai at about 1:50 a.m., and then crossing the Bohai Sea, the Yellow Sea, and the north of the East China Sea on the following two days.

4.2. MTSAT-Derived Dust Properties Mapping and Comparison

The retrieval algorithm mention above was applied to the detected dust pixels. Figure 4 shows one of the Streamer simulation results when the cloud temperature and surface temperature were 260 K and 285 K, respectively.

Figure 4. The BT₁₁ and BT₁₁ − BT₁₂ for individual pixel values (red dots) of MTSAT data on the Streamer simulation tables.

Figure 4. The BT₁₁ and BT₁₁ − BT₁₂ for individual pixel values (red dots) of MTSAT data on the Streamer simulation tables.

Any individual red dot is determined by the BT₁₁ and BT₁₁ − BT₁₂ observed from MTSAT. The black curves and the blue curves are the BT₁₁ and BT₁₁ − BT₁₂ simulated from Streamer model at different effective radii and AOTs. They are used to generate a lookup table. It can be seen that BT₁₁ is almost decreasing linearly with increasing optical thickness, while BT₁₁ − BT₁₂ decreases with an almost linear increase of particle effective radius. Since the observation of satellite is the mixture of the surface and dust cloud temperature, under a fix surface temperature, the lower BT₁₁ usually means the higher altitude the dust cloud has reached. The higher altitude may correspond to heavy dust aerosol loading with higher AOT. Depending on this physical pattern, the optical thickness and particle effective radii of dust particles can be retrieved according to the position of the red dots in the grids.

Then, the dust column density can be obtained using the derived equations. The dust density may be different from different source and transport routes. The specific density for Sahara dust is reported as ρ = 2.6g/cm³ [27]. For Asian dust, however, this value is usually taken as 2.5 g/cm³ [16,23], which was also used in our study. Figure 5a–c shows the MTSAT retrieved results of the dust optical thickness, effective radius and column density at 2:30 a.m. on 2 March 2008. Unfortunately, the retrieval results lack robust validations with the in situ measurements. To further increase confidence on the performance of our algorithm, the results are compared with the MODIS retrieval at 2:05 a.m. (Figure 5d,f).

Figure 5. Comparison between MTSAT (a–c) and MODIS (d–f), retrieved at nearly simultaneous time points; (a) MTSAT AOT; (b) MTSAT effective radius; (c) MTSAT column dust density(gm⁻²); (d) MODIS AOT; (e) MODIS effective radius; (f) MODIS column dust density(gm⁻²).

Figure 5. Comparison between MTSAT (a–c) and MODIS (d–f), retrieved at nearly simultaneous time points; (a) MTSAT AOT; (b) MTSAT effective radius; (c) MTSAT column dust density(gm⁻²); (d) MODIS AOT; (e) MODIS effective radius; (f) MODIS column dust density(gm⁻²).

Figure 6. Correlation plots of the dust column concentrations obtained from MODIS compared to MTSAT.

Figure 6. Correlation plots of the dust column concentrations obtained from MODIS compared to MTSAT.

Focusing on the major regions of the dust cloud, the MTSAT retrieved AOT is largely consistent with the MODIS. However, for some regions, especially the outer rim, it shows relatively large differences. This is mainly due to the rapid variations that occur in the outer rim within a short time. Moreover, the MODIS and MTSAT retrievals should not be expected to precisely the same. The dust column density has a generally similar pattern and good correlation between MODIS and MTSAT retrievals (Figure 6).

4.3. Estimation of the Total Deposited Mass

The satellite observation is a snapshot of one scenario of the dust event. It is a comprehensive reflection of dust transport, diffusion and deposition. It is difficult to tell these processes apart from each other from the satellite. Thus, the variation of the space average of the dust column density is used to derive the deposition ratio. The horizontal transport velocities are usually much larger than the deposition velocities. To minimize the influence of horizontal transport, only the space average of the dust column density, in which a standard deviation less than a threshold is selected to build the regression model. The threshold is set to 1.5g/m² from experience. The average dust column density $\overline{{\text{M}}_{\text{du}}}$ decreases exponentially with an increase of dust transport time (Figure 7). Thus, an exponential model was created from the selected $\overline{{\text{M}}_{\text{du}}}$. The regression model is shown as follows:

y=ae ^−bx + c

(8)

where y is the value of mean dust column density, x is time from the dust began to enter the seas, and a, b, and c are 5.509, 0.142 and 0.35, respectively, for this case. These are regressed from the model, and the correlation coefficient between x and y is R² = 0.925. This empirical model may be limited for this case study in the Bohai Sea and the Yellow Sea and for other cases, the coefficients a, b, and c of the model y=a × e ^−bx + c can to be regressed by our method. Further study is needed to improve the model and to make it more robust. The ratio of deposition ($\overline{v_{\text{du}}}$) was determined by the derivative of the exponential model

$$\overline{{\upsilon }_{\text{du}[x]}}=-0.782e^{-0.142x}$$

(9)

Figure 7. Variation of dust deposition rates (g/m²) with relative time from when the dust reached the Bohai Sea.

Figure 7. Variation of dust deposition rates (g/m²) with relative time from when the dust reached the Bohai Sea.

The deposition amount M_de for each pixel is derived using Equations (6) and (7) and the space distribution is shown in Figure 8.

Figure 8. The spatial distribution of dust depositon during the period of 1–2 March 2008.

Figure 8. The spatial distribution of dust depositon during the period of 1–2 March 2008.

The coarser dust particles are removed faster than the finer ones. Therefore, total deposition flux decreased rapidly on the way from the Bohai Sea (~20 g/m²) to the Yellow Sea (~6 g/m²). Table 3 shows the annual deposition flux in the study region from previous cruise and site measurements and modeling results at different time periods. The values are vary.

Table 3. The dust deposition over the Bohai Sea and the Yellow Sea (g/m²·year).

Authors	Time Period	Methods	The Bohai Sea	The Yellow Sea
Liu and Zhou [28]	1987–1992	cruise track	26.4	9.3
Uematsu et al. [29]	1994–1995	model	17.5	9.7
Gao et al. [30]	1989	sites	30~196	8.6~79

Table 3. The dust deposition over the Bohai Sea and the Yellow Sea (g/m²·year).

Authors	Time Period	Methods	The Bohai Sea	The Yellow Sea
Liu and Zhou [28]	1987–1992	cruise track	26.4	9.3
Uematsu et al. [29]	1994–1995	model	17.5	9.7
Gao et al. [30]	1989	sites	30~196	8.6~79

Uematsu et al. [31] used four island stations to estimate the total deposition of atmospheric mineral to the central North Pacific to be ~20 Mt·yr ⁻¹ from January 1981 to March 1982. Gao et al. [32] estimated the total atmospheric deposition of dust to the China Sea as about 67 Mt·yr ⁻¹ for 1992. Kai and Huiwang [33] estimated the total amount of dry deposition to the Yellow Sea to be 17.9 Mt·yr ⁻¹ from the difference between the inland city (Beijing) and the coastal city (Qingdao) samples during 2000–2002. There were five dust-transporting events in this study region based on the national sand-dust weather almanac and dust detection procedure for 2008. The total deposition amount into the Bohai Sea and the Yellow Sea during the case event was estimated to be 2.59 Mt. If the total deposition amount of this case event were multiplied by five, the magnitude would be roughly comparable to previous in situ and model results. The values are varied because the intensity and duration of the dust events change greatly over time and space as well as the estimating methods were different.

5. Conclusions

This study provides an alternative method to estimate the deposition flux of dust aerosol over the seas from geostationary satellites. The MTSAT observations are used to investigate a dust event over the Bohai Sea and the Yellow Sea during 1–3 March 2008. The major conclusions are as follows: (1) Streamer is suitable to simulate the brightness temperature and the simulations can be used to create the LUT under different conditions; (2) spatial mapping of retrievals of the dust clouds with MTSAT agrees with nearly synchronous MODIS products and the correlation of dust column concentration reaches 0.85; (3) the dust column concentration decreased exponentially with the dust transport time on the pathway; and (4) the magnitude of the deposition flux is comparable to the in situ and model results from previous studies.

This paper demonstrates that two-band MTSAT retrievals of dust events could be used to routinely quantify the deposition of dust storm events over the seas under reasonable assumptions. These methods can be used to create a long-term time series of dust deposition flux in the coastal seas of China. The satellite estimated deposition flux would be a direct index for investigating the biogeochemical implications in this area for further study.