The Spatiotemporal Patterns of Surface Water Temperature In a Brazilian Hydroelectric Reservoir

1

The spatiotemporal patterns of surface water temperature in a Brazilian hydroelectric reservoir

Enner Herenio Alcântara José Luiz Stech

João Antônio Lorenzzetti Evlyn Márcia Leão de Moraes Novo

Brazilian Institute for Space Research, Remote Sensing Division

E-mails: {enner, stech, loren, evlyn}@dsr.inpe.br Abstract The water temperature plays an important role in the ecological functioning and controlling the biogeochemical processes of a water body. Conventional water quality monitoring is expensive and time consuming. Particularly problematic if the water bodies to be examined are large. Conventional techniques also bring about a high probability of undersampling. Conversely, remote sensing is a powerful tool to assess aquatic systems. Based on this, the objective of this study was to map the surface water temperature and improve understanding of spatiotemporal variations in a hydroelectric reservoir. In this work the MODIS land-surface temperature (LST) level 2, 1-Km nominal resolution data (MOD11L2, version 5) was used. All available clear-sky MODIS/Terra imagery between 2003 and 2008 were used, resulting in a total of 786 daytime and 473 nighttime images. Descriptive statistics (mean, maximum and minimum) was computed for the historical images, so as to build a time series of daytime and nighttime monthly mean temperature. The thermal amplitude and the anomaly were also computed. In-situ meteorological variables were used from 2003 to 2008 to help us understand the spatiotemporal variability of the surface water temperature. The surface energy budget and the depth the wind can distribute a given surface heat input were also measured. A correlation between daytime and nighttime surface water temperature and the meteorological parameters and a linear regression computed. These relationship and the causes of the spatiotemporal variability was discussed. Keywords: Water surface temperature; heat flux; mixed depth layer; thermal amplitude. Introduction Reservoirs, or man-made lakes, are usually built to store water for later use, water supply,

flood control or power generation (Casamitjana et al., 2003). In Brazil, there are

approximately 31 hydroelectric reservoir buildings by electric sector with volume more than 1

billion of m3. The hydroelectric sector is responsible for 97% of energy generation and

considered the largest hydroelectric park of the world (Kelman et al. 2002). The building of

these dams, however, causes greatest environmental, social and economics impacts (Tundisi,

1994). Over the time, the functioning of reservoirs affects its retention time. Ford (1990)

INPE ePrint: sid.inpe.br/mtc-m18@80/2009/09.29.19.24 v1 2009-09-30

2

comparing lakes and reservoirs with the same morphometry concluded that inflow and

withdrawal lead to a decrease in the reservoir’s water retention time.

In accordance to Kimmel et al (1990) to understanding the performance and functioning of

reservoir ecosystems the water temperature distribution is fundamental. Residence time

affects the temperature distribution in the water column. Therefore, a deeper understand on

the aquatic environment functioning taking depends on a better assessment of the temperature

distribution in three dimensions depth, space and time.

Surface water temperature is a key parameter in the physics of aquatic system processes since

it accounts for surface-atmosphere interactions and energy fluxes between the atmosphere and

the water surface. As it influences water chemistry it also affects its biological processes

(Lerman and Imboden, 1995). Moreover, temperature differences between water and air

control moisture and heat exchange in the air/water boundary and as a consequence, are

crucial for understanding the hydrological cycle.

Conventional water quality monitoring is expensive and time consuming. Particularly

problematic if the water bodies to be examined are large. Conventional techniques also bring

about a high probability of undersampling. Conversely, remote sensing is a powerful tool to

assess aquatic systems and is particularly useful in remote areas (Alcântara et al. 2009).

Thermal infrared remote sensing applied to freshwater ecosystem has aimed to map surface

temperature (Oesch et al., 2008; Reinart and Reinhold, 2008; Crosman and Horel, 2009), bulk

temperature (Thiemann and Schiller, 2003), circulation patterns (Schladow et al., 2004) and to

characterize upwelling events (Steissberg et al., 2005). However, the application of thermal

infrared image to study surface water temperature in hydroelectric reservoir is scarce and, in

Brazil it is being attempt for the first time.

Several satellites have been launched with spatial, temporal and radiometric resolutions for

study surface water temperature with relatively accuracy (Steissberg et al. 2005). However,

most of them acquire data twice at every 16 days, as Landsat and ASTER (Advanced

Spaceborne Thermal Emission and Reflection Radiometer) at a given location. And the

Landsat satellite does not record data at night (unless by special request). The Moderate

Resolution Imaging Spectroradiometer (MODIS) on board of Terra and Aqua satellites

overcome these temporal resolution limitations. MODIS data can typically be acquired daily

due to the large scan angle (Justice et al., 1998).

Though, the objectives of this study are to map the surface water temperature in the Itumbiara

hydroelectric reservoir and to improve the understanding of its spatial and temporal patterns.


3

The Itumbiara hydroelectric reservoir (18°25’S, 49°06’W) is located in a region stretched

between Minas Gerais and Goiás States (Central Brazil) originally covered by tropical

grassland savanna. Damming the Parnaiba River flooded backward its main tributaries:

Araguari and Corumbá River. The basin geomorphology resulted in a lake with a dentritic

pattern covering an area of approximately 814 Km² and volume of 17.03m³ (Figure 1).

(a) (b)

(c)

Figure 1: Localization of Itumbiara hydroelectric reservoir on Brazil’s central (a), on state context (b) and on regional scale (c) with the bathymetric map. On regional scale is showing the flooded area over a SRTM (Shuttle Radar Topography Mission) image.

The reservoir was built in 1979 and started its operation in 1980. Figure 2 shows the reservoir

area before the flooding (Figure 2-a) and after (Figure 2-b) flooding.


4

(a)

(b)

Figure 2: (a) MSS-Landsat-3 imagery from 11-08-1978 show the area before inundation; and (b) TM-Landsat-5 imagery from 26-05-2007 actual period. The figure also shows the position of dam on reservoir.

The climate in the region is characterized by precipitation ranges from 2.0 mm in the dry

season (May - September) to 315 mm in rainy season (October - April). In the rainy season

the wind intensity ranges from 1.6 to 2.0 ms-1 and reaches up to 3.0 ms-1 in the dry season

(Figure 3-a). The air temperature in the rainy season ranges from 25 to 26.5 ºC and breaks

down to 21ºC in June as the dry season starts. The relative humidity has a pattern similar to

that of the air temperature, but with a little shift of the minimum value towards September

(47%). Moreover during the rainy season the humidity can reach 80% (see Figure 3-b).

(a) (b)

Figure 3: Climate patterns on Itumbiara reservoir: mean monthly of (a) precipitation (mm month-1) and wind intensity (m s-1), (b) air temperature (ºC) and humidity (%).

The following section will describe the methodology needed to explain the spatiotemporal

variability in the water surface temperature.


5

Methodological approach

The methodological approach was developed using the concept of disturbing influences that

the reservoirs is exposed and described by Fischer et al. (1979) as: (1) the meteorological

conditions in the area will determine the strengths of any energy transfers across the air-water

interface; (2) the water from the inflowing streams may impart kinetic and potential energy,

and (3) some of the energy of the outflowing water may be transformed to kinetic energy of

the reservoir water.

Satellite data

MODIS land-surface temperature (LST) level 2, 1-km nominal resolution data (MOD11L2,

version 5) was obtained from the National Aeronautics and Space Administration Land

Processes Distributed Active Archive Center (Wan, 2008). All available clear-sky MODIS

Terra imagery between 2003 and 2008 were used, resulting in a total of 786 daytime images

and 473 nighttime images (Figure 4).

Figure 4: Acquisition date and time of all MODIS/Terra data for 2003-2008 used in this study.

Maps of monthly mean daytime and nighttime surface water temperature were produced for

the period from 2003 to 2008. Descriptive statistics (global mean, maximum and minimum)

was computed for the surface water temperature maps so as to build a time series of daytime


6

and nighttime monthly mean temperature. The thermal amplitude was computed pixel-by-

pixel by subtraction daytime and nighttime temperature.

To obtain the anomaly the monthly mean temperature for the years spanning from 2003 to

2004 was computed in a pixel bases and then subtracted from each month for the entire time

interval. The seasonal thermal amplitude was also analyzed.

In situ Data

Daily mean air temperature (ºC), relative humidity (%), wind intensity (ms-1) and

precipitation (mm) from 2003 to 2008 were used. These data was obtained from a

meteorological station (see Figure 1 for location). The daily mean of each variable was

converted into monthly mean to adequate it to the time scale of satellite data. The surface

energy budget was also calculated to help the understanding the surface water temperature

variability.

Surface Energy Budget

The study of exchanges of energy between lake and atmosphere is essential for the

understanding the aquatic system behavior and its reaction to possible changes of

environment and climatic conditions (Gianniou et al., 2007). The exchange of heat across the

water surface was computed using the methodology described by Henderson-Sellers (1986)

as:

radroLriSsN AA φφφφφ −−−+−= )1()1( (1)

Where Nφ the net energy available, sφ is the incident short-wave radiation, riφ the incident

long-wave radiation, SA and LA is short-wave and long-wave reflectivities (albedoes), roφ

the (blackbody) long-wave radiative loss, and radφ the net non-radiative energy loss (sensible

and latent heat flux). The units used for the terms in Eq. (1) are W m-2.

Moreover, the energy exchange also occurs through precipitation, withdrawal of evaporated

water, chemical and biological reactions in the water body, and conversion of kinetic to

thermal energy. These energy terms are small enough to be omitted. Many researchers agree

that omitting the energy budget components with small values does not significantly affect the

results (Bolsenga, 1975; Sturrock et al., 1992; Winter et al., 2003). The sensible and latent

heat flux will be calculated for daytime and nighttime using the monthly mean surface water

temperature derived from MODIS data.


7

As pointed out by Ford and Johnson (1986), the wind is the major source of energy for many

physical phenomena which both directly or indirectly cause mixing in the water system and

also can alter the thermal patterns. So, the depth to which wind can mix the water reservoir

will be calculated to help in the results discussion.

Mixed Depth

Using the wind intensity from meteorological station and the net heat flux calculated by

surface energy budget (Eq. 1) the depth (tD , m) to which wind can mix the reservoir is given

by (Sundaram, 1973):

p

Nk

t

CgB

wD

ρφα

3*= (2)

Where *w is the shear velocity of the wind (ms-1), Bk an empirical coefficient approximately

equal to von Karman´s constant (0.4), α the volumetric coefficient of thermal expansion for

water (1.8x10-4 ºC-1), Nφ the surface heat flux (Wm-2), ρ the density of water (≈ 1000

kgm-3), g is gravitational acceleration (9.8 ms-2), and pC the specific heat of water (4186 J

kg-1 ºC-1).

The physical significance of tD is that it is a measure of the depth the wind can distribute a

given surface heat input. If the aquatic systems have a depth greater than tD and not

dominated by advection, this system will probably stratify (Ford and Johnson, 1986).

To try to explain which meteorological parameters best explain the surface water temperature

variation, a statistical analysis will be computed.

Statistical Analysis

Pearson’s correlation and regression analysis were applied to relate the daytime and nighttime

surface water temperature derived from MODIS image against the meteorological variables

(Neter, 1996).

Results and Discussion

The figure 5 shows the monthly mean daytime and nighttime surface water temperature

distributions at Itumbiara reservoir. In general way the daytime temperature decrease from

boundary to center of the reservoir and for the nighttime the processes invert. The daytime


8

temperature in May is more homogeneous than the others and for nighttime the months of

July and August presents this homogeneity.

Figure 5: Monthly mean of daytime and nighttime surface water temperature from 2003 to 2008.

The daytime series shows that January is the month with the smallest Maximum temperature,

which starts to rise in February. From March to July the maximum temperature rises and is

kept stable around ± 28ºC. From August to October the temperature raises around 7ºC in


9

relation to July. In November and December the temperature starts to drop. The mean

temperature presents the same patterns observed in the maximum temperature.

The daytime time series shows that the maximum temperature is coldest during January with

a little rising in February (Figure 6-a). Between March and July the maximum temperature is

little variable. Between August and October a rising of temperature occurs (~ 7ºC) in relation

to July. The months of November and December presents a little down in maxima

temperature. The mean temperature presents the same patters observed in the maxima

temperatures. The minimum temperatures present an expressive down during March and May

and after July the minimum temperature stats to rising.

(a)

(b)

Figure 6: Monthly mean surface water temperature statistics from 2003-2008.

The monthly mean nighttime temperature presents a small variability during the years than

daytime (Figure 6-b). Their pattern of variation is well defined with highest temperature in

January and May and October and December. The smallest temperature occurs between July

and September. The comparison between daytime and nighttime temperatures reveals that the

highest temperatures occur in September for daytime and January for nighttime whereas the

lowest ones occur in May and July, respectively.

The figure 7 shows the daytime and nighttime mean temperature difference on a monthly

base. January has the smallest temperature difference from day to night (<1ºC). From January

on the mean temperature difference increases and reaches its maximum in October (8.46ºC).


10

Figure 7: Surface water mean temperature difference between daytime and nighttime from 2003 to 2008.

Interanual Temperature Anomaly

The interanual temperature anomaly was analyzed and shows that the January presents the

highest anomaly between 2003 and 2008 for daytime temperature; following by September,

November and December (Figure 8-a). The smallest anomaly was observed for June. The

histogram of the anomaly for daytime shows the occurrence of the more frequently

temperature anomaly of ± 2ºC (Figure 8-b).

The same anomaly pattern was observed for daytime and nighttime surface water temperature,

but with March presents the highest anomaly variability in the years of study (Figure 8-c). In

this case the variability is small than observed during daytime. The histogram of the anomaly

for nighttime shows the occurrence of the more frequently temperature anomaly of ± 1ºC

(Figure 8-d).


11

(a)

(b)

(c)

(d)

Figure 8: Interanual-mean temperature anomaly and histogram distributions for daytime (a, b) and nighttime (c, d) respectively.

Seasonal surface water temperature

The analysis of the seasonal changes of surface water temperature shows that the thermal

amplitude between daytime and nighttime is negative for summer in great area of the

reservoir (Figure 9). This means that the nighttime temperature is highest than daytime

temperature during summer. This means that the nighttime temperature is higher than daytime

temperature during the summer. The highest differences occur in the center of the reservoir

(~-6ºC).

In autumn the thermal amplitude is near zero with t negative amplitudes occurring in the

central portion of the reservoir. In winter this negative thermal amplitude does not occur and

are replaced by patches of near zero amplitudes in the central portion of the reservoir. In the

spring, these patches of near zero amplitude are smaller with the occurrence of positive

differences (Figure 9).


12

Summer

Autumn

Winter

Spring

Temperature (ºC)

Figure 9: Mean-surface water temperature difference between daytime and nighttime of stations of the year from 2003 to 2008.

Thus, in general form it is possible to verify that a trend exist with positively differences from

border to central water body of the reservoir from summer to spring. It is due to low depth in

these areas, less than 1 m. During the spring highest positively temperature differences can

occur (~ 6ºC). The nighttime temperature could be highest than daytime during summer and

autumn. However, during the summer this is more pronounced than autumn. To

understanding these variability observed in the results the energy fluxes was computed.

Surface energy budget for daytime and nighttime

The daytime sensible heat flux is negative in January, September and October (Figure 10-a).

For the nighttime is negative from August to December (Figure 10-b). The negative sensible

heat flux occurs when the surface losses heat by convective and advective processes. When

the flux is positive the surface gain heat.

The latent heat flux was positive to all months during daytime and nighttime. During January,

February and May the latent flux for nighttime was higher than daytime. This occurs because

the loss of heat by surface during nighttime was more high than daytime.


13

(a)

(b)

Figure 10: Energy budget components for (a) daytime and (b) nighttime.

The daytime net flux was always positive whereas the nighttime was always negative. This

happens because during daytime the primary source of the energy to warm up the water

surface is the incoming shortwave radiation and the main source of loss of heat is the

evaporation. During nighttime this source of short wave radiation does not exist and all

storage energy during the day will be exchanged through the atmosphere. The wind is also a

great source of energy that could modify the surface water temperature through the water

column mixing due to surface water wind action.

Daytime and nighttime mixed depth

During the daytime the mixed depth is high in January, August, October and December and

low during February to May and moderate in June, July and November. The highest mixed

depth can occurs during September when the mixed depth can reach 3.8 m (Figure 11).


14

Figure 11: Annual wind mix depth (m) at Itumbiara reservoir.

The mixed depths during the nighttime are lower than during daytime and the highest mixed

depth occurs during July (2.4 m). This mixed depth came to down from October to December

and starts to rise from January to June.

It is clear that during the daytime a combination of incoming short wave radiation and wind

blowing in the surface water is responsible to distribute the heat inside the water column and

from August to October a combination of high air temperature and wind intensity and low

precipitation and relatively humidity (see Figure 3) contribute to cause a deepest heat

distribution in the water column. In the beginning of the year the depth which the wind can

mix is low, less than 2 m during daytime and 0.5 during nighttime.

On the other hand, during the nighttime don’t have contribution of incoming short wave

radiation, but we have a lost of heat (gained during the day) trough the exchange between

surface water and atmosphere and the action of the wind blowing.

Statistical model for daytime and nighttime surface water temperature

The Pearson correction computed between daytime and nighttime surface water temperature

derived from the MODIS image against meteorological parameters are present at Table 1. For

the daytime temperature the significant correlated meteorological parameters are air

temperature, evaporation, short wave radiation and long wave radiation. For nighttime


15

temperature the meteorological correlated parameters are air temperature, humidity, wind

intensity, precipitation, evaporation, sensible heat flux and net heat flux.

Table 1: Pearson correlation coefficients for daytime and nighttime surface temperature against air temperature (Tair), humidity (H) wind intensity (W), precipitation (P), evaporation (Ev), short wave radiation (SW), long wave radiation (LW), sensible flux ( Sφ ), latent flux ( Lφ )

and net flux ( Nφ ).

Daytime water temperature Nighttime water temperature

Tair 0.82 0.73

H - 0,78

W - -0.75

P - 0.85

Ev 0.83 0.72

SW 0.65 -

LW 0.77 -

Sφ - 0.82

Lφ - -

Nφ - -0.86

Only significant values at 95% significance level are shown.

The multiple regression analysis shows that for daytime surface water temperature the

correlated meteorological parameters explain 92% of the annual variation

(RMS=0.62ºC, 0001.0=ρ ). For nighttime the meteorological parameters explain 94%

(RMS=0.45ºC, 0006.0=ρ ). The representative equations of these relationships are presented

in equations 3 and 4.

)72.0()96.0()33.4()2.3(49.137 LWSWEvTairdaytime +−++−+−= (3)

+−+++= )26.0()052,0()16.2(78.25 WHTairnighttime

)40.0()35.0()08.1()28.0( SNEvP φφ +−+−+

(4)

Where: Tair is the air temperature (ºC), H is humidity (%),W is wind intensity (ms-1), P the

precipitation (mm),Ev is the evaporation (mm), SW is the short wave radiation (W m-2),

LW is the long wave radiation (W m-2), Nφ is the net flux and Sφ is the sensible flux.


16

The daytime surface water temperature driven forces as shown in equation 3 is related to

heating processes though the short wave, and consequently heating of the air temperature.

Then the lake emits thermal radiation (outgoing long wave) and evaporates the surface water.

Because of this the daytime surface water temperature following the air temperature (see

Figure 3-b) and evaporation signals (see Figure 10-a).

The nighttime surface water temperature pattern is more complex than daytime because it

depends on multiple parameters. The air temperature continues important to explain the water

variability but the precipitation and latent and sensible flux is important also. The

precipitation is important because during the raining season the difference between daytime

and nighttime temperature is lower than during the dry season (see Figure 7).

The wind intensity explain the cooling phases of the surface temperature because during the

occurrence of a lake breezes (from land to water) the reservoir loss heat (Arrit, 1987;

Simpson, 1994); this is especially true during September when the wind intensity is higher

than others months of the year (see Figure 3-a). The equations 3 and 4 will help us to discuss

the results shown above.

As shown in the equation 3 and at table 1 the temporal variability found in the daytime

surface water temperature is due mainly by evaporation and air temperature. In fact the water

temperature (Figure 6-a) follows the same pattern found in the air temperature (Figure 3-b).

This is obviously due to heat transfer from atmosphere to the water. Moreover, the

evaporation processes is responsible for cooling the water surface, causing a differential heat

and cooling mechanisms.

This mechanism is responsible for the interanual anomaly observed for daytime (Figure 8-a).

That is, during January when the anomaly is higher than others month the sensible heat flux is

negative (heat losses) and this is variable from year to year. However the incoming short

wave radiation is high and consequently the air temperature also. The heat losses occur due to

precipitation that in this month is high. The variability of the precipitation should be a major

source of the interanual anomaly observed in the daytime water temperature. This is

supported also due to the fact that when the precipitation breaks the interanual anomaly drops

down until 1± ºC.

During the nighttime when the water surface just losses the heat, the main source of anomaly

is the amount of heat stored during the day and the differential heat and cooling mechanism.

The effect of the precipitation regime for nighttime temperature is shown in the equation 4

and the correlation analysis (R2 = 0.85) at table 1. For this, the interanual-mean temperature

anomaly for nighttime is lower than during daytime (Figure 8-b and 8-d).


17

Summary and Conclusion

The objective of this study was to map the surface water temperature and improve

understanding of spatial and temporal variations in the Itumbiara hydroelectric reservoir. The

main conclusions are:

- The spatial variation of both daytime and nighttime surface water temperature follows the

air temperature signal;

- Surface temperature time variation during daytime is higher through the year as compared to

that of nighttime;

- The thermal amplitude is more pronounceable during September and October and the lowest

amplitude occurs in January;

- The interanual mean anomaly is larger during January and September and smaller during

June for both daytime and nighttime;

- The negative amplitude occurs mainly in summer season and heating from land to the center

of the reservoir;

- September is the month when the latent flux is higher than the others months of the year and

the sensible flux is negative booth for daytime and nighttime;

- The highest depth mix caused by wind intensity in daytime occurs in September and in July

for nighttime;

- The driven forces for daytime surface temperature are air temperature, evaporation, short

and long wave radiation;

- For nighttime are the air temperature, humidity, wind intensity, precipitation, evaporation,

sensible and latent flux.

Acknowledgment The authors would like to thank the FAPESP Project 2007/08103-2.

References Alcântara, E.; Barbosa, C.; Stech, J.; Novo, E. & Shimabukuro, Y. (2009). Improving the

spectral unmixing algorithm to map water turbidity distributions. Environmental Modelling & Software, 24, 1051-1061.

Arrit, R.W. (1987) The effect of water surface temperature on lake breezes and thermal

internal boundary layers. Boundary-Layer Meteorology, 40, 101-125. Armengol, J.; Garcia, J.C.; Comerma, M.; Romero, M.; Dolz, J.; Roura, M.; Han, B.H.; Vidal,

A. & Simek, K. (1999). Longitudinal processes in canyon type reservoirs: the case of Sau


18

(N.E. Spain). In Tundisi, J.G. & Straskraba M. (eds), Theoretical Reservoir Ecology and its Applications: Backhuys Publishers. Leiden. The Nederlands: 313-345.

Bolsenga, S. (1975). Estimating energy budget components to determine Lake Huron

evaporation. Water Resources Research, 11, 661-666. Casamitjana, X.; Serra, T.; Colomer, J.; Baserba, C. & Pérez-Losada, J. (2003). Effects of the

water withdrawal in the stratification patterns of a reservoir. Hydrobiologia. 504: 21-28. Crosman, E.T. & Horel, J.D. (2009). MODIS-derived surface temperature of the Great Salt

Lake. Remote Sensing of Environment, 113, 73-81. Fischer, H.B.; List, E.J.; Koh, R.C.Y.; Imberger, J.; Brooks, N.H. (1979). Mixing in inland

and coastal waters. Academic Press: California. Ford, D.E. Reservoir transport process. In: Thorton, K.W..; Kimmel, B.L.; Payne, F.E. (ed.).

Reservoir limnology. Ecological Perspectives. John Wiley and Sons. New York: 15-41. 1990.

Ford, D.E. & Johnson, L.S. (1986). An assessment of reservoir mixing processes. Technical

Report E-86-7, U.S. Army Engineers Waterways Experiment Station, Vicksburg, MS. Henderson-Sellers, B. (1986). Calculating the Surface Energy Balance for Lake and Reservoir

Modeling: A Review. Reviews of Geophysics, 24, 625-649. Iqbal, M. (1983). An Introduction to Solar Radiation. Library of Congress Cataloging in

Publication Data. Academic Press Canadian. Justice, C.O.; Vermote, E.; Townshed, J.R.G.; Defries, R.; Roy, D.P.; Hall, D.K.;

Salomonson, V.V.; Privette, J.L.; Riggs, G.; Strahler, A.; Lucht, W.; Myneni, R.B.; Knyazikhin, Y.; Running, S.W.; Nemani, R.R.; Wan, Z.; Huete, A.R.; Van Leeuwen, W.; Wolfe, R.E.; Giglio, L.; Muller, Jp.; Lewis, P.; Barnsley, M.J. (1998). The moderate Resolution Imaging Spectroradiometer (MODIS): land remote sensing for global change research. IEEE Transactions on Geoscience and Remote Sensing, 36, 1228-1247.

Kelman, J., Pereira, M.V.F., Neto, T.A.A., Sales, P.R.H. Hidreletricidade. In: Rebouças, A.C.;

Braga, B.; Tundisi, J.G. (eds.) Águas Doces no Brasil. São Paulo: escrituras, 2002. pp. 371-418.

Kimmel, B.L.; Lind, O.T.; Paulson, L.J. Reservoir primary production. In: Thorton, K.W..;

Kimmel, B.L.; Payne, F.E. (ed.). Reservoir limnology. Ecological Perspectives. John Wiley and Sons. New York: 133-194. 1990.

Lerman, A., Imboden, D.; Gat, J. 1995. Physics and chemistry of lakes. Springer-Verlag,

Berlin. 334 p. Neter, J.; Kutner, M.H.; Nachtsheim, C.J.; Wasserman, W. Applied linear statistical models. 4

ed. Boston: McGraw-Hill, 1996.


19

Oesch, D.; Jaquet, J.-M.; Klaus, R. & Schenker, P. (2008). Multi-scale thermal pattern monitoring of a large lake (Lake Geneva) using a multi-sensor approach. International Journal of Remote Sensing, 29, 5785-5808.

Reinart, A. & Reinhold, M. (2008). Mapping surface temperature in large lakes with MODIS

data. Remote Sensing of Environment, 112, 603-611. Schladow, S.G.; Palmarsson, S.O.; Steissberg, T.E.; Hook, S.J. & Prata, F.J. (2004). An

extraordinary upwelling event in a deep thermally stratified lake. Geophysical Research Letters, 31, L15504.

Simpson, J.E. (1994). Sea breeze and local wind. Cambridge University Press, Cambridge.

234p. Steissberg, T.E.; Hook, S.J. & Schladow, S.G. (2005). Characterizing partial upwellings and

surface circulation at Lake Tahoe, California-Nevada, USA with thermal infrared images. Remote Sensing of Environment, 99, 2-15.

Sturrock, A.; Winter, T. & Rosenberry, D. (1992) Energy budget evaporation from Williams

Lake: a closed lake in north central Minnesota. Water Resources Research, 28, 1605-1617.

Sundaram, T.R. (1973). A theoretical model for seasonal thermocline cycle of deep temperate

lakes. Proc. 16th Conf. on Great Lakes Res., 1009-1025. Thiemann, S. & Schiller, H. (2003). Determination of the bulk temperature from

NOAA/AVHRR satellite data in a midlatitude lake. International Journal of Applied Earth Observation and Geoinformation, 4, 339-349.

Tundisi, J.G. Tropical South America: presents and perspectives: In: Margalef, R (ed)

Limnology now: a paradigm of planetary problems. Amsterdan: Elsevier Science. 1994, pp. 353-424.

Wan, Z. (2008). New refinements and validation of the MODIS land-surface

temperature/emissivity products. Remote Sensing of Environment, 112, 59-74. Winter, T.; Buso, D.; Rosenberry, D.; Likens, G.; Sturrock Jr., A.; & Mau, D. (2003).

Evaporation determined by the energy-budget method for Mirror Lake, New Hampshire. Limnology and Oceanography, 48, 995-1009.


The Spatiotemporal Patterns of Surface Water Temperature In a Brazilian Hydroelectric Reservoir

Documents