4. DISEÑO DE LA PLANTA DE TRATAMIENTO DE AGUA POTABLE
4.3. OTROS TANQUES PARA EL PROCESO
Passive microwave sensors measure radiation emitted from the surface of the earth as brightness temperatures. The components contributing to the magnitude recorded include the upwelling atmospheric emission, earth’s surface emission attenuated by atmosphere, atmospheric downwelling atmospheric emission reflected at surface and attenuated along the upward path by the atmosphere and cosmic background emission attenuated by atmosphere reflected at the surface and attenuate again along the upward path by the atmosphere (Wigneron et al. 2017). Passive sensors have a high revisit period e.g. Soil Moisture and Ocean Salinity (SMOS) and Soil Moisture Active Passive (SMAP) 2-3 days. Communications and broadcast systems cause Radio Frequency Interference (RFI) resulting in abnormally high brightness temperatures. RFI signals affect brightness temperatures at a higher intensity in the C band as compared to the X band (Oliva et al. 2012). Radio frequency interference has led to degraded measurements in some parts of Europe, South Asia, Middle East and China.
124
The RFI has reduced through sensitization in concerned countries to protect the 1400-1427 MHz SMOS operation frequency (Oliva et al. 2012). Radiometer measurements are most accurate in areas of low vegetation. However, lower frequencies have higher penetrative ability suitable for measurements in vegetated areas. The penetration ability of L band radiometers makes it an asset in the retrieval of soil moisture. Additionally, L band has a deeper sampling depth, approximately 0-3 cm as compared to C band 0-1 cm and are affected less by atmospheric effects (Wigneron et al. 2017). Unique values are assigned to areas with water, ice snow, excess surface roughness and steep topography. Algorithms in L band soil moisture retrieval include:
1) 2-Parameter Retrieval Algorithm: SMOS soil moisture is based on this algorithm. Multi-angular dual polarized observations of brightness temperature are inverted using L Band Microwave Emission of the Biosphere (L-MEB) model to obtain soil moisture and optical depth at nadir. Incidence angles range from 0-55° with brightness temperatures for the ‘optimal location’ viewed from different angles. The value at the optimal location is used as a first approximation in the 2-parameter retrieval of soil moisture and optical depth at nadir. No estimates are required in the retrieval.
2) Single Channel Algorithm (SCA): SMAP is based on this algorithm. TB is measured at one incidence angle, 40°, at one polarization. TB are corrected for vegetation and roughness and finally, a dielectric mixing model is used to retrieve soil moisture.
3) Dual Channel Algorithm: This is an extension of the SCA where both polarizations, H, and V from SMAP are combined to retrieve soil moisture by minimizing the RMSE between simulated and observed Temperature Brightness (TB).
4) Land Parameter Retrieval Model (LPRM): Microwave Polarization Difference Index (MDPI) is computed from the TB and the 𝜏 − 𝜔 model used to retrieve soil moisture. In densely vegetated areas, LPRM models do not converge and thus no data is recorded.
5) Multi-Temporal Dual Channel Algorithm: Combined consecutive TB observations taken in the early morning when the temporal changes in Vegetation Water Content (VWC) are constant are used in retrieval of soil moisture. The retrieved albedo is constant due to the stability of the surface temperature in early hours of the morning.
125 Table 6.1: A summary of the data used
Soil Moisture products Spatial resolution Temporal resolution Format Period assessed Data portal
FLDAS_NOAH 0.1° Daily .NC 2010-2016 https://ldas.gsfc.nasa.gov/FLDAS/FLDA Sdownloadphp.
FLDAS_VIC 0.25° Daily .NC 2010-2016 https://ldas.gsfc.nasa.gov/FLDAS/FLDA Sdownloadphp.
ERA-Interim 0.125° Daily .NC 2010-2016 http://apps.ECMWF.int/datasets/data/Int erim-full-daily/levtype=sfc/ SMOS 30-50 km Daily .NC 14/1/2011 -2016 https://SMOS- diss.eo.esa.int/socat/SMOS_Open/search Vegetation product Spatial resolution Temporal resolution Format Period assessed Data portal
MODIS 250 m 16-day .HDF 2010-2016 https://lpdaac.usgs.gov/dataset_discovery /modis/modis_products_table/mod13q Rainfall product Spatial resolution Temporal resolution Format Period assessed Data portal TRMM_3B42 0.25° Daily .NC 2010-2016 https://disc2.gesdisc.eosdis.nasa.gov/data/ TRMM_L3/TRMM_3B42_Daily.7/ 6.2.1.1 SMOS
It was launched in 2009 has a spatial resolution of 30-50 km. Ascending mode in the early morning minimizes perturbation due to air, vegetation and soil temperature in the L band (Kerr et al. 2001). Moreover, L band has a high sensitivity to changes in soil moisture and salinity in the ocean and has a higher penetration depth up to 5 cm which performs well in dense vegetation (Kerr et al. 2001). On the other hand, high-frequency products have a penetration depth of 1-2 cm preventing penetration into the soil layer in vegetated environments. SMOS has multi-angular dual polarization capabilities hence any points on the surface are viewed frequently at different angles and polarizations. The angular information separates the different contributions to the signal. The angular signature retrieves soil moisture and vegetation optical depth (vegetation water content) which expresses the quantity of signal that is absorbed by the vegetation layer through minimization of a cost function between L band microwave emission of biosphere model (L-MEB) and the corresponding SMOS measured temperature brightness to estimate soil moisture and VOD. L-MEB is associated with certain land cover classes thus making it possible to quantify the contribution of these classes (Kerr et al. 2012). The retrieval algorithm consists of
126
static (soil texture from FAO, ECOCLIMAP land use maps, topography index) and dynamic (rain, temperature, snow from ECMWF) datasets. Radio Frequency Interferences, vegetation opacity and surface roughness have regular updates for future inversions.
HQN model is used for roughness parametrization. For low vegetation, hr is 0.1, for forests, it is set as 0.3, Q is 0 whereas Nv=0 and Nh=2. Dobson model was used in modeling dielectric constant before L2 v5.5. Currently, Mironov model is in use. Mironov improves retrievals over dry warm surfaces reducing extreme values in soil moisture. ECMWF determines the vegetation and soil temperature. 𝜏 − 𝜔 model used to model vegetation. Albedo for low vegetation set to 1 while that for forests is set to 0.06-0.08. The vegetation structure is assumed to be isotropic i.e. tth=ttv=1. Optical depth at nadir and soil moisture are byproducts of the 2 parameter retrievals by inversion of the LMEB model. Optical depth at nadir is modeled by a linear function of LAI.
The spatial extent of SMOS (on average, 40x40 km²) poses a challenge due to the introduction of significant spatial variability in soil moisture. Moreover, due to the spherical nature of the earth, the view angle changes therefore different areas covered at different angles (Kerr et al. 2012). Additionally, SMOS surface soil moisture retrieval maps are available though there are gaps associated with RFI, steep topography, dense vegetation, snow cover, frozen soils. Data products available are Level 1(TB) and level 2(ocean salinity over oceans or soil moisture/vegetation opacity over land) and level 3 (1d, 3d, 10d or particular month for the globe either morning or afternoon passes for soil moisture and vegetation opacity). To improve spatial resolution, SMOS data can be combined with other data such as MODIS data.
Albedo describes the fraction of solar energy reflected from the earth back into space. It is also optical brightness scaled from 0 to 1. The ratio of irradiance reflected to irradiance received by a surface. E.g. Ice albedo 0.5 to 0.7 means 50 to 70%of the incoming radiation is reflected. While for ocean 0.060 means only 6% of incoming solar radiation is reflected and the rest is absorbed.
Reflectivity describes the light reflected from the surface in relation to light incident upon it. The fraction of radiant energy reflected from a surface.
Emissivity is the ratio of energy radiated from a surface in relation to that radiated from a black body at the same temperature and wavelength under the same viewing angles.
The 𝜏 − 𝜔 model computes emission from a two-layer soil water medium in H and V polarization using three terms; upwelling emission by vegetation, downward emission from the vegetation reflected by soil and attenuated by vegetation canopy and emission by soil attenuated by canopy (Wigneron et al. 2017). The terms account for influence of vegetation and soil roughness on angular measurements and polarization mixing. The temperature brightness is given by:
127 𝑇𝐵 = 𝑇𝑠𝑜𝑖𝑙(1 − 𝑅𝑠𝑜𝑖𝑙,𝑝)𝑉𝐴𝑝+ (1 − 𝜔𝑝)𝑇𝑣𝑒𝑔(1 − 𝑉𝐴𝑝) + (1 − 𝜔𝑝)𝑇𝑣𝑒𝑔(1 − 𝑉𝐴𝑝)𝑅𝑠𝑜𝑖𝑙,𝑝 𝑉𝐴𝑝= exp (−𝜏𝑝⁄cos 𝜃) 𝑇𝑠𝑜𝑖𝑙 𝑖𝑠 𝑒𝑓𝑓𝑒𝑐𝑡𝑖𝑣𝑒 𝑡𝑒𝑚𝑝 𝑜𝑓 𝑠𝑜𝑖𝑙, 𝑇𝑣𝑒𝑔 𝑖𝑠 𝑒𝑓𝑓𝑒𝑐𝑡𝑖𝑣𝑒 𝑡𝑒𝑚𝑝 𝑜𝑓 𝑣𝑒𝑔, 𝑅𝑠𝑜𝑖𝑙,𝑝 𝑖𝑠 𝑠𝑜𝑖𝑙 𝑠𝑢𝑟𝑓𝑎𝑐𝑒 𝑟𝑒𝑓𝑙𝑒𝑐𝑡𝑖𝑣𝑖𝑡𝑦, 𝜏𝑝 𝑖𝑠 𝑣𝑒𝑔 𝑜𝑝𝑡𝑖𝑐𝑎𝑙 𝑡ℎ𝑖𝑐𝑘𝑛𝑒𝑠𝑠, 𝜔𝑝 𝑖𝑠 𝑎𝑙𝑏𝑒𝑑𝑜 𝑜𝑓 𝑣𝑒𝑔 𝑐𝑎𝑛𝑜𝑝𝑦, 𝑉𝐴𝑝 𝑖𝑠 𝑣𝑒𝑔 𝑎𝑡𝑡𝑒𝑛𝑢𝑎𝑡𝑖𝑜𝑛 𝑓𝑎𝑐𝑡𝑜𝑟 Roughness modeling
The HQN model is used to model roughness.
𝑅𝑠𝑜𝑖𝑙,𝑝= [(1 − 𝑄)𝑅𝑝+ 𝑄𝑅𝑞]𝑒𝑥𝑝−𝐻𝑝cos𝑁𝑝𝜃
𝑝 = 𝐻 𝑤ℎ𝑖𝑙𝑒 𝑞 = 𝑉 𝑝𝑜𝑙𝑎𝑟𝑖𝑧𝑎𝑡𝑖𝑜𝑛 , 𝑅𝑝& 𝑅𝑞 𝑎𝑟𝑒 𝑡ℎ𝑒 𝑟𝑒𝑓𝑙𝑒𝑐𝑡𝑖𝑣𝑖𝑡𝑦 𝑜𝑓 𝑎 𝑠𝑚𝑜𝑜𝑡ℎ 𝑠𝑜𝑖𝑙 𝑠𝑢𝑟𝑓𝑎𝑐𝑒, 𝐻𝑝 𝑖𝑠 𝑡ℎ𝑒 𝑟𝑚𝑠ℎ 𝑜𝑟 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦 𝑜𝑓 𝑟𝑜𝑢𝑔ℎ𝑛𝑒𝑠𝑠, 𝑄 𝑖𝑠 𝑡ℎ𝑒 𝑝𝑜𝑙𝑎𝑟𝑖𝑧𝑎𝑡𝑖𝑜𝑛 𝑚𝑖𝑥𝑖𝑛𝑔 𝑝𝑎𝑟𝑎𝑚𝑒𝑡𝑒𝑟,
𝑁𝑝 𝑎𝑙𝑙𝑜𝑤𝑠 𝑓𝑜𝑟 𝑑𝑒𝑝𝑒𝑛𝑑𝑒𝑛𝑐𝑒 𝑜𝑓 𝑟𝑜𝑢𝑔ℎ𝑛𝑒𝑠𝑠 𝑜𝑛 𝜃 𝑄 = 0, 𝑁𝑝= 0 𝑎𝑛𝑑 𝑁𝑝 = 2 ℎ𝑒𝑛𝑐𝑒 𝑅𝑠𝑜𝑖𝑙,𝑝 = 𝑅𝑝𝑒𝑥𝑝−𝐻𝑝cos𝑁𝑝𝜃
Roughness intensity increases with increasing roughness, Q, the difference between emissivity in horizontal and vertical polarization decreases as Q increases whereas Q shows an increase with increasing frequency.
Vegetation modeling
The vegetation optical thickness (height of vegetation canopy), is a function of optical thickness at nadir, incidence angle, and polarization (CBSA et al. 2014).
𝜏𝑝= 𝜏𝑁𝐴𝐷(sin2𝜃 𝑡𝑡
𝑝+ cos2𝜃) 𝑝 = 𝐻, 𝑉
A value of tt>1 results to increasing 𝜏𝑝 as a function of 𝜃. In cases where 𝜏𝑝 is independent of polarization and incidence angle, tt=1. The parameter tt accounts for effect of vegetation anisotropy on optical depth with respect to incidence angle, polarization. A good relationship was found between height of stems and tt v. *property of being directionally dependent, which implies different properties in different directions