1
Evapotranspiration (ET) Lab
February 8, 2016
Estimation of ET from Multiple Methods
Overview
This lab has two objectives:
1)Compares estimates of forest evapotranspiration from several methods.
2)Estimate the effect of forest practices (harvesting, planting density) on a stand level water budget.
Introduction
Rainfall is partitioned onto several pathways after falling. First, it may be intercepted by the “architecture” of whatever ecosystem it falls on. That may be the leaves, stems and branches of a forest, or the roads and buildings of urban areas. This interception can be between 10 and 50% of the water budget, depending of factors like the amount and intensity of rainfall, and the type and characteristics of the vegetation. This water is generally evaporated back to the atmosphere, but may also be stored or used. Second, rainfall may fall through the canopy, either directly (throughfall) or via stems (stemflow). This water can either enter the soil where it can be taken up by roots and transpired or runoff over the land surface. The character of this partition is dependent, as before, on the amount and intensity of the rainfall, on the infiltration capacity of the soil (which can be dramatically modified by heavy machinery and other land cover changes) and by the amount of rainfall that has occurred prior to the present (so-called antecedent rainfall). Any water that makes it into the soil starts to fill the pore spaces and, if possible, percolate downwards. As water passes through the soil, plants can use it (transpiration) or it can be returned to the atmosphere abiotically (evaporation).
The sum of the interception that is evaporated, the evaporation of throughfall, and the transpiration of soil water is evapotranspiration, and it is generally the biggest water loss component, often by far. The word evapotranspiration implies that it is a coupled process that combines both biotic and abiotic mechanisms. While there are techniques to disentangle the components, it is generally modeled as one quantity, though one that is sensitive to the vegetation at the site where it’s estimated.
Among the most important insight about the ET process is that it requires a LOT of energy. Remember the latent heat of vaporization? That’s the energy required to convert liquid water at 100 deg C to vapor water at the same temperature. It takes 2.45 MJ (million Joules, where the Joule is the SI unit of energy) to do that for 1 kg of water. For comparison, the energy in a gallon of gasoline is about 120 MJ; that is 1 gallon of gas can evaporate only roughly 13 gallons of water. This value (the latent heat of vaporization) is a function of temperature, but we’ll treat it as a constant here (2.45 MJ/kg).
It’s also critical to remember that air has a capacity to hold water that is a function of temperature of that air. Specifically, the es (saturation vapor pressure, measured in kPa, a measure of pressure) is predicted by the following equation:
[1]
Where T is temperature in degrees C, and es is in units of kPa. The vapor deficit is important, so is the slope of the saturation vapor pressure line –we use both. We compute the slope based on the following equation:
[2]
Where es is from the equation above, T is temperature (in Celsius) and Δ is the slope in units of kPa per deg C. We’ll compute both equations in class to determine their values as a function of temperature ourselves. It’s worth pointing out that the energy of radiation in (from the sun) is lost in two main ways: latent heat export (i.e., in the energy required for the phase conversion from liquid to vapor form) and sensible heat flow (i.e., the energy associated with a change in air temperature). The ratio of those two pathways (latent vs. sensible heat fluxes) is an important component of an ecosystems energy budget; it is called the Bowen Ratio.
A final item that is necessary to compute is the psychrometric constant (γ), which for our purposes has a constant value of 0.0667 kPa °C-1.
The other aspect of the lab is to understand the effects of stand management on ET. Two papers are provided that describe ET processes in production stands as a function of stand age, and in naturally regenerated stands. Differences between these imply that production forestry may use a lot of water; that is, even though forests aren’t irrigated in North Florida, their water use exceeds what a conserved forest would use, sometimes by a lot. Your task will be to use those two papers to understand the magnitude of that difference, as well as better understand the ET process.
TASK 1
ET is estimated using methods that rely on temperature alone, those that rely on radiation, and those that consider a combination of factors (e.g., temperature, relative humidity, radiation). We also have methods that measure the process directly, without relying on atmospheric proxy variables. They vary in complexity, and in accuracy. Our task is to compare them.
Compute ET using three methods:
a)McCloud method (1955)
b)Hargreaves method (1985)
c)Penman method (1963)
McCloud Method
This is among the simplest methods for estimating ET, and is based on temperature alone. The equation is:
[3]
Where ETp is potential evapotranspiration (that is, what would occur if water were not limiting, as is designed to be the case in irrigated agriculture, for which this equation was developed) in mm per day, K is a conversion parameter (K = 0.254), W is a parameter that describes the sensitivity of ET to temperature (W = 1.07) and T is the mean daily temperature in degrees F.
1) Compute ET using McCloud for both 2001 and 2008 data provided. Plot both vs. Julian day (where Jan 1 = 1, Jan 2 = 2,…, Dec 31 = 365).
Hargreaves Method
The Hargreaves method is an empirical model that was developed for forage systems in California. However, its simplicity and use of readily available data make it desirable to employ elsewhere. It depends on solar radiation and mean air temperature data only.
[4]
Where Rs is the solar radiation reaching the Earth’s surface (measured in MJ/m2/d), λ is the latent heat of vaporization (a constant equal to 2.45 MJ/kg) and T is mean daily temperature (this time in degrees C). ET is computed in mm/day.
2) Compute ET using Hargreaves for both 2001 and 2008 data provided. Be careful of units! Plot both vs. Julian day (where Jan 1 = 1, Jan 2 = 2,…, Dec 31 = 365).
Penman Method
The Penman method is the gold-standard for evapotranspiration, incorporating considerations of the radiation at a site, and the effects of air movement. It is pretty complicated, but breaks down into two parts, both which depend on computations made above.
[5]
Where ET is the evaporation in mm/day, λ is the latent heat of vaporization (2.45 MJ/kg), Δ is the gradient of the saturation deficit (computed in Eq. 2), γ is the psychrometric (assumed constant at 0.0667 kPa °C-1), Rn is the net radiation in MJ/m2/d, G is the soil heat flux (which we will neglect here), T is the temperature in °C, U is the average windspeed at 2m aloft in m/s, and D is the saturation vapor deficit (es – ed). The saturation pressure (es in units of kPa, computed in Eq. 1) is a direct function of temperature. The actual vapor pressure (ed, also in units of kPa) is the saturation pressure multiplied by the relative humidity (RH, in units of %). That is, ed = RH * es. All the data necessary to compute this has been provided. Don’t forget to divide the right hand side of the equation by 2.45 MJ/kg at the end to get ET in mm/day.
3)Compute ET using Penman for both 2001 and 2008 data provided. Be careful of units! Plot both vs. Julian day (where Jan 1 = 1, Jan 2 = 2,…, Dec 31 = 365).
4)How viable is the assumption that there is no soil storage of energy (that is, that G is zero)? This can be evaluated qualitatively by looking at the changes in soil temperature over time. Plot soil temp vs. time. If the curve is not a straight line then soil temperature varies systematically with season.
5)Plot measured ET vs. estimated ET for each of the methods (McCloud, Hargreaves, Penman) for 2001 and 2008; there will be 3 graphs. Fit a line to the data that is forced through the origin (0,0), and report the slope and r2. Which model appears to overestimate measured ET? Underestimate? Which relationship is the most noisy?
6)Plot a curve of cumulative measured ET and measured rainfall vs. time for both 2001 and 2008. What is implied by the location of the two lines? During 2001, how much of the year was in cumulative rainfall surplus? What about 2008? Based on these two graphs, what period of year would you predict forests to be MOST water stressed?
7)[BONUS +2] Explore the data in the “White Method Data” tab. Use the White Method to calculate ET for the period or record:
- Plot the time series of water level over the period of record. What do you see as a general trend? What about within-day variation?
- Calculate ET = Sy * (H + S), where ET is in cm/d, Sy is 1.0 (unitless), H is the slope of the water table at night (i.e., between midnight and 5 am) in units of cm/day and S is the water elevation change between midnight on day a and midnight on day a+1 (units of cm/day).
- Do the rates make sense? Are they roughly constant over the period of record?
TASK 2
Three papersare attached (Gholz and Clark 2002, Powell et al. 2005, McLaughlin et al. 2013). The first two describe two sets of measurements of evapotranspiration, while the third is a synthesis of stand-level water use in the southeast, focusing on the impacts of forest management on water yield. The first paper looks across a chronosequence of forest stands from recently clear cut to rotation aged. The second measures ET from a low-density naturally-regenerated mixed pine forest in Austin Cary Memorial Forest. While the data are for a dry period (1998 and 1999 for the first paper, 2000 through 2002 for the second) they still several key insights in the paper about the effects of stand density management on water use. Your task is to read all three papers and answer the following questions:
From the Gholz and Clark (2002) paper:
1)How does the Bowen Ratio change with stand age (Table 5)? What does this mean for the temperatures in the vicinity of a clear cut vs. a mature stand?
2)What do the data in Table 7 imply about the effects of stand development on ET and interception? What does it mean when they report a value of ET + I (evapotranspiration + interception) greater than precipitation?
3)Assuming ET stabilizes when leaf area has maxed out (around mid-rotation thinning), what fraction of the rainfall is used? Compare the daily ET values in winter and summer (from Table 6) with values in Powell et al. (2005) paper (Table 4) for a conserved forest.
4)Can you think of any other way to measure the hydrologic effects of stand management (particularly clear cut harvest) other than measuring ET directly?
From the Powell et al. (2005) paper:
5)Compare the attributes between the rotation aged stand in Gholz and Clark (2002) and the naturally regenerated stand in Powell et al. (2005). What are the major differences?
6)The paper suggests naturally regenerating stands use what fraction of annual rainfall? What factors (beyond the authors control) might dramatically affect that percentage?
7)What is implied by Figure 5?
8)From Table 4, compute the annual ET difference between a mature production stand and a naturally regenerated low density flatwoods (assume 180 days of summer, 180 days of winter). What is the magnitude of additional ET (in mm/yr) expected from a watershed converted from low-density to production pine? Where is that water coming from?
From the McLaughlin et al. (2013) paper
9)What is water yield? Why would small changes in ET from a forest stand potentially dramatically affect water yield?
10)What is the most important stand attribute for predicting ET? How do you measure that attribute?
11)What are the simulated effects of fire and clear-cuts? Why is the comparison of an intensive silviculture site in 20-25 year rotations with a conserved forest not fair?
12)[BONUS +1] Assuming you use 150 gal/d of water, how many acres of dense pine stand (~20 m2/ha) would need to be thinned to a basal area of 14 m2/ha to offset your water consumption?