Tarija Bolivia Climate Part 2 V5b

D R A F T V5b
Notes on Climate Data, Watershed Characteristics, and the Relationship of Precipitation, Infiltration, and
Evapotranspiration for Water Balance Studies in Tarija, Bolivia
David Stolpa, Docente UCB, January, 2019
The purpose and importance of water balance studies

The history of civilization is, in a sense, the history of the relationship between mankind and water. The
relationship is so critical that “water security” is now one of the many terms being used to describe the future of
many of the world’s regions. Water quantity, free of contamination and of a quality to be meet the needs of both
increasing populations and the environment, is recognized as an essential element for survival. Water balance
studies are a part of the planning process for assuring that sustainable supplies can be developed or supplemented.
Part 1: El Valle Central
The study region

The City of Tarija (elevation 1854 m [1], Lat S 21o 32' 48", Long W 64o, 42' 39" at the airport) is the capital of the
of the Department of Tarija in Bolivia. Bolivian departments are equivalent to a US state. The department consists
of four climatic/geomorphic zones. The City of Tarija is located in Zone 3 which is also known as El Valle
Central.
Figure 1: The four zones of the Department of Tarija. The City of

Tarija and El Valle Central are located in Zone 3. El Valle Central
(about 3333 km2 +) is drained principally by the Guadalquivir River
and its two major tributaries; the Camacho and Santa Ana Rivers
(author)
El Valle Central is the focus of political and economic activity in the department and contains the highest
concentration of the population. It is also the center of the greatest growth within the department (about 390,000 -
Instituto Nacional de Estadistica). The primary industry in the department is gas production. The primary industry
in EL Valle Central is agriculture.
Climatic setting
Zone 3 includes distinct rainy and dry seasons. The rainy season lasts about 7 months and generally occurs from
October to April (late spring through early autumn in the southern hemisphere). The dry season lasts about 5 months
from May through September (mid-autumn to mid-spring).
Fortunately, the wet season also coincides with the growing season, thus helping to reduce the reliance on irrigation.
However, the rains can be erratic and many small farmers without access to irrigation must wait for the rains to
plant. Harvests can be at risk without irrigation.
Annual average precipitation ranges from 490 mm to 1200 mm per year depending on the specific location in El
Valle Central, and the years of record under analysis. Daily temperatures in the vicinity of the City vary from 8o to
30o C. Hard freezes resulting in icy conditions at ground level in the valleys are rare. Snow and ice in the
mountains can, and do occur but infrequently and last for only short durations. Frost in the higher elevations is
more common. Micro climates are also relatively common. However, the overall region can basically be
described as semiarid. Regional humidity averages 59 percent. The dominant vegetation cover is scrub land.
[1] These note use the convention of the comma “,” for metric values, and the decimal point “ .” for US Customary units.
1
D R A F T V5b
Climate data, precipitation, and streamflow gages are scarce and often consist of a short record of less than 10
years. Many gages have been discontinued. Streamflow gages when they exist, they are usually limited to gaging
only channel flows and not flood flows. High bed mobility and deep, unconsolidated beds results in uncertainty in
the readings. Recording gages for climatic data and precipitation have only been recently introduced.
Thus, the major challenge in defining the climate and modeling the hydrologic response of watersheds in the
Valle Central is the lack of climatic and hydrologic data. Climatic data for the Department is available from
“Servicio Nacional de Meteorología e Hidrología” (SENAMHI). Note that much of the climate observations in
this paper are mostly generalities due to the scarcity of climatic stations and detailed records.
An informal trend analysis of precipitation by the author indicates that the existence of the usual cycles of short-
term wet and dry periods are common. However, no conclusive long-term climatic trends are obvious. Generally,
the variation that does occur (such as reduced streamflow) is commonly blamed on climate change. The more
likely cause for the decrease in streamflow is the increased population and associated diversions. Again, more
study is needed.
A second informal analysis indicates no definite correlation between the El Niño and La Niña effects of the
Pacific Ocean to the west (560 km). This is likely due to the high altitude of the principal range of the Andes
Mountains which serve as a barrier for advection. The other older and lower subranges to the east also serve as a
barrier to moisture from the Atlantic Ocean. The moisture that does arrive in the region is from the Atlantic.
However, it must travel in a round-about way to Argentina to the south/southeast where El Valle Central is more
open. It then moves up the valley.
Figure 2: The hills and mountains to the east and west of El Valle
Central obstruct the direct advection of moist air from the Atlantic
and Pacific Oceans. (author)
Geological setting
The geology of El Valle Central is interesting. The hills and mountains (peaks as high as 14,000 feet to the west in
the Sama range) exist to either side of the north-south trending valley. They consist of parent rock of marine
sediments that are highly faulted, distorted and fractured as a result of the tectonic and associated mountain
building process that created the Andes Mountains. The parent rock consists of shales, siltstones, and sandstones
(GEOBOL 1991). The oldest formations are of the Cambrian age (more than 485 million years). Next, are
formations associated with the Ordovician and possibly Silurian epochs (420 to 485 million years). Finally, there
are the formations that are associated with the Devonian period (360 to 430 million years). The rocky areas of the
hills and mountains account for over 50 percent of the surface area of El Valle Central.
The more recent geological character of El Valle Central began to develop during the Last Glacier Maximum
(44,000 to 21,000 before the present). The area was much wetter during this period and experienced a series of
advancing and retreating glaciers and associated glacier lakes. Erosion and glacier action resulted in deep deposits
of lacustine sediments having an average depth of 50 m + and reportedly as much as 500 m. The sediments
consists of gravels and cobble, clays, silts, and sands. Included with the lacustine deposits are layers of volcanic
ash. The region is famous for being one of the richest Pleistocene fossil sites in South America (Clarke).
2
D R A F T V5b
Part 2: Theories, concepts, and issues
About Data
Both the quality and quantity of reliable data are important factors in the development and use of
hydrologic/hydraulic models.
Figure 3: General climatic data for Tarija

The description of the available climatic data for Tarija can vary even when provided by the same source
(Figure 3: Weatherbase in this case). Data can depend on several factors. Spatial and time scale are the general
bounding conditions. The date of development, protocol at time of collection, variation in instrumentation,
failure of instrumentation, improvements in technology, calibration of instrumentation, siting of instruments,
and purpose of intended use are factors among many others. Finally, there is the problem of interpretation of the
data, climatic/watershed processes, and concepts by the modelers. All will impact the reliability of hydrologic
studies.
Figure 4: Different siting conditions for a pan evaporator

can impact recorded values. (FAO 56)
Figure 5: Two similar cross-sections with the same gage

depth. However, B actually carries more water because of
the undetected underflow in the deeper beds. If the flow
conditions are assumed to be equal for A and B based
solely on the gage readings, an error will be introduced
into the streamflow analysis. Case B is typical of many
streams in the Department of Tarija (author)
Figure 6: A 24-hour storm spanning two days. The 24-hour depth is

30 mm. However, the first day rainfall is only 6 mm. The second day
rainfall is 24 mm. Although 24 hours and one (1) day are often
synonymous, there is only one storm and two different methods for
recording the values. The difference in description can have significant
impact on interpretation and analyses.(author)
Sometimes the only option about data is to do the best that we can.
3
D R A F T V5b
About Hydrologic Models
There are two general types of hydrologic models:
• Statistical (gage analyses and regression equations), and
• Deterministic (rainfall-runoff and water balance models)
Statistical: Statistical models are based on random data that occur naturally. The data is evaluated by an
application to an assumed distribution that describes the probability of an outcome. An analysis of streamflow
data that changes randomly and naturally from year to year will result in different values for a particular sub-
set of specific years under analysis. The latest analysis with new data is usually considered to be better than
the analysis of the previous years. Log Pearson Type III streamflow analysis is an example of a statistical
models.
If the annual flows are changed by irrigation diversions due to an increase in users or an introduction of
reservoir control into the watershed, the assumed distribution may not apply. In such cases a different
statistical approach such as a trend analysis to partition the natural data epochs from the controlled epochs
may be needed. The change from natural to controlled conditions due to the rapid increase in population of El
Valle Central is an important consideration for the analysis of streamflow data for Tarija.
Deterministic: Deterministic models utilize input parameters and details “determined” by the modeler. Inputs
can include fixed sets of statistical results that the modeler may occasionally update. However, as long as the
inputs are the same each time the model is performed, the outcomes will always be the same. Rainfall-runoff
models such as the Rational equation and unit hydrograph methods are examples of deterministic models.
Model detail is an important characteristic. “Lumped” models may include general watershed conditions and
hydraulic methods such as a channel routing procedure for considering flow attenuation. Lumped models
often serve as a good, simple start if data or time to complete a more detailed analysis is lacking. As
additional or detailed data become available, or more time is made available, the original lumped model can
and should be improved.
The various unit hydrograph (such as Snyder’s) approaches for rainfall-runoff modeling are an example of a
more detailed approach than an application of the Rational equation. The lack of detail can limit the
application of a model. The Rational equation is generally limited to an area of about 80 hectares (200 acres).
Inputs can also consist of a mix of sub-sets of empirical data or approaches that are more conceptual or
theoretical. The Green-Ampt infiltration method is an example of a conceptual model preferred by academics.
The Natural Resources Conservation Service (NRCS) runoff curve number (RCN) method is an example of an
empirical model preferred by practical modelers. The NRCS was formerly known as the Soil Conservation
Service (SCS, the name change occurred in the mid 1990’s).
A total discussion of hydrologic models can be complex and is beyond the scope of this document. However,
several simple rules (adapted from the American Society of Civil Engineers or “ASCE” standards for
evapotranspiration or ET methods) should be considered when selecting a water balance method:
• The method and equations should be understandable
• The method, equations, and concepts should be defensible and should be traceable to quality field
measurements. For example, it is difficult to justify using ET for a site with a surface dominated by rock
and very little soil. The moisture capacity of intact rock essentially does not exist.
• The approach should use accepted methods.
• The approach should maximize simplification without significant loss of accuracy.
• The approach should use existing, readily available data.
Models should be verified using historical data or by using complementary methods. A typical complementary
approach for streamflow is to use a statistical approach to check a deterministic approach.
4
D R A F T V5b
Water balance models
Two categories (and two schools of thought) of water balance model applications exist. Both are deterministic but
also use statistical data. These are:
• Models based on watershed characteristics (principally infiltration), that are used for determining water
availability for agricultural, industrial, municipal, and domestic use. A special case is for planning
reservoir storage.
• Climatically-based models (evapotranspiration – ET) for determining soil water budgets and schedules
used for irrigation
Water availability: The objective of water availability models is to determine runoff (such as streamflow) that
can be diverted or stored for future use. The general equation takes the form of:
Q=P–I (1)
Q = Runoff
P = Precipitation such as rainfall and snow, and
I = Infiltration (including initial abstraction Ia), ET is also often used
Additional detail or parameters can be added to the equation such as diversions for municipal use.
Soil water budgets for irrigation: The objective of these studies is to determine soil moisture availability for
irrigation needs. The amount of soil moisture available is usually evaluated as a function of evaporation and
plant transpiration. Evaporation (E) is the process where moisture from surface water of soil water enters the
atmosphere driven by climatic processes. Transpiration (T) is the process by which plant roots intake moisture
from the soil (osmosis) and release it through stomata in the leaves. Separation of the two processes can be
complex and so they are usually lumped together as a single process commonly referred to as
evapotranspiration (ET). Note that there are no universally accepted standard terms in the practice although
the need for standardization has been often discussed. The equations for irrigation applications takes the
general form of:
Available Soil Moisture = (Precipitation, irrigation, ground water, or a combination) – ET (2a)
Soil moisture can be measured through the use of lysimeters or estimated ET equations.
Note that Equation 1 and 2 are conceptually equal. That is, the amount of available water for runoff or plant use is
equal to a “loss” measurement subtracted from a water source value. Therefore, ET has sometimes been used as
surrogate for infiltration for water availability models such that:
Q = P – ET (2b)
Infiltration can be the dominating loss term in the equations for watershed yield computations. ET will usually be
the dominating term for irrigation applications. The two loss terms, “I” and “ET” can be included in the same
equation under specific circumstances. For example the Water Erosion Prediction Project or WEPP hillslope
model uses the equation:
Θ = Θin + (P−I) + S − Q − ET − D – Qd
Per a rearrangement of the terms in Equations 2a and 2b, “I” and “ET” have
also been considered to be interchangeable (I = ET) under specific
circumstance and often only one or the other will appear in an water
availability equation. Consideration of both processes in basin water
availability have their respective advantages and disadvantages. However,
infiltration and ET are two completely different processes in the hydrologic
cycle and depend on different categories of input data with some overlap.
Figure 7: The accumulated vapor in a plastic bag enclosing a plant is the

result of transpiration. (Ming Ket College)
Infiltration in a model is primarily dependent on; 1, characteristic watershed soil type; 2, soil antecedent moisture
conditions (AMC now referred to as antecedent runoff conditions ARC); 3, vegetation and land-use cover; and 4,
precipitation intensity. There are other conditions.
5
D R A F T V5b
The only direct climatic input is precipitation. AMC is a climatic function of ET but it is only marginally treated
in a climatic manner, usually in terms of rainfall interevent time. Vegetation type is also climatically-related but is
treated in general descriptive terms in infiltration models. Infiltration models can also account for changes in the
watershed.
Some examples of “infiltration” conditions include:
• Sandy watershed soils will have more infiltration and less runoff that clayey watershed soils
• More runoff will occur when soils are wet (high AMC)
• Areas with poor grassy covers will generally have more runoff than those of heavily wooded areas
• Runoff only occurs when initial abstractions have been satisfied and when rainfall intensities exceed
infiltration rates
Tarija generally lacks information on soil type, infiltration rates, and sub-24 hour rainfall intensities. Soil AMC,
while not formally investigated is almost always dry based on simple observations.
Evapotranspiration models are mostly dependent on climatic factors. They can have precipitation and vegetation
type inputs in common with infiltration-based models but the two inputs for both models can be evaluated
differently. ET models can present significant uncertainty when watershed conditions vary considerably from
those that were used to develop the specific ET approach or when watershed conditions are highly variable or are
changing.
Part 3: Precipitation (P), infiltration (I), and evapotranspiration (ET)

Precipitation (and irrigation water) is the source of water for ET. The
amount of precipitation “P” that replenishes soil moisture depends
primarily on infiltration “I” and to a lesser extent on the presence of
ground water.
Figure 8: Ground water is sometimes a source of soil moisture.

(USGS)
Precipitation “P”
Precipitation is a function of storm type. Common descriptions for storm types that produce rainfall are small
convective storms, meso-scale convective complexes (derechos), frontal, stratiform, orographic, and tropical
cyclonic lows (including hurricanes). Common types in Tarija are convective, frontal, and orographic.
The duration and depth of a precipitation event depends the on several conditions. Included are the triggering
conditions for a rainfall event to occur such as dew point, on how much moisture the local atmosphere can hold,
and on how fast the atmospheric moisture can be replenished as a storm event progresses. Proximity to a source of
moisture (oceans or even a large lake systems) is important. However, the presence of all three conditions is
essential. The Atacama Dessert is adjacent to the Pacific Ocean (source of water) but the triggering and other
conditions are problematic (constant temperature inversion, moisture advection blocked by mountains, etc.,)
All rainfall events vary in time and space. A time distribution of rainfall intensity is called a hyetograph. Rainfall
intensity relative to a specific time period is important in evaluating infiltration and runoff.
A wide variety of precipitation gages are used collect rainfall data. The rainfall depth recorded by a gage is
usually considered to be a “point” value. A point value is assumed to represents a uniform depth over an area of
25 km2 (10 square miles). For general modeling purposes the gage value adjusted for storm size area is usually
considered to occur uniformly over a watershed. In explanation, the assumption of rainfall uniformity breaks
down as storm and watershed areas increase over 25 km2. Gaps can occur and intensities can vary between the
storm cells occurring over a larger area. Therefore, point rainfalls are usually adjusted (reduced) by an areal
adjustment factor.
6
D R A F T V5b
Figure 9: Two rainfall events having different areal

coverage captured by a gage as having a depth of
20 mm. However, the two storms result in different
runoff conditions (author).
One issue is that the recorded “point” rainfall depth by a gage at any given time may not be representative of what
is actually happening within a watershed. Given sufficient gage coverage and time, the recorded data may
approach an average representative value. The possibility of using Thiessen polygons might be considered to
adjust data for coverage. However, El Valle Central watershed with its various sub basins and micro climates is
really too complicated for the method to work well. A topographic divide that defines an abrupt change in local
climatic conditions may not be captured by a conventionally delineated polygon.
Precipitation data is available in many forms. For single event applications, the annual maximum values can be
converted into depth-duration-frequency (DDF) values or intensity-duration-frequency values. Duration can be
measured in minutes, hours, and even days.
Other forms of precipitation data can include annual totals for a specific year and annual averages over several
years. The same concepts of totals and averages can be applied to other time increments such as daily, weekly,
monthly, and even seasonal values.
Generally, smaller (sub 24 hour) time increments and intensities are used when evaluating infiltration for a single
event. Larger time increments (daily, weekly, monthly, etc.) are used for considering infiltration for water balance
studies.
The concept of “effective” precipitation and streamflow

The magnitude or volume of “usable” streamflow is usually that which is contained within the channel. Flow that
exceeds the channel capacity is considered flooding and except for some incidental capture in reservoirs, it
normally passes downstream and is lost. Thus, bank-full discharge represents an upper limit of water that is
“effectively” available. The concept has its equivalent in irrigation practices where there is just so much water that
can be used by the plants. Extra irrigation water is wasted.
The concept is often ignored in water availability studies and thus can result in an over estimation of the supply
that can actually be applied to meet demand. One excuse for ignoring the limit is that streamflow is not always at
bank full discharge and that flooding magnitudes will offset the partial, low-flow conditions that are the “norm”.
This is not necessarily a good assumption and the data should be evaluated to determine the potential for bias
when flooding flows are included above a usable limit.
Channel capacity is the result of the channel-forming flow. The channel-forming flow is a function of the
magnitude and frequency (sometimes modal and not necessarily a median value) of the events primarily
responsible for the “stable” geometry of the channel. It is usually estimated to be a magnitude slightly less than
that of the 50 percent annual exceedance probably (AEP, 2 - year return period) flow. One approach is not to
entirely discard extreme precipitation events but simply to reduce the extreme event depth to that of the two-year
precipitation value.
As a simple example, the average annual rainfall for Tarija has sometimes
been expressed as 490 mm. The 100-year, 24 -hour rainfall can be 110
mm. If a 100-year precipitation value is included in the annual average,
the amount of the usable water available will likely be overestimated.
Figure 10: The significance of modal values. Eleven of the 12 months
for an example region have rainfall depths of 11 mm. One month
includes an extreme event and registers 120 mm. The annual total is
241 mm. The “average” annual depth of 20 mm is normally assumed
to be representative. It is not in this case. Engineering judgement must
be used if a representative value is to be applied in an analysis (author).
7
D R A F T V5b
Soil moisture
On average, soils consist of 50 percent solids of which 45 percent is mineral and 5 percent organic material.
About 50 percent of the soil volume consists of voids of which half is occupied by water and half by air and other
gases. The composition and structure of soils can vary considerably due to soil type, clast size, disturbance such
as compaction.
Figure 11: Composition of soils. Relative density (specific

gravity) of soil can range from 1.1 to 1.6 (author)
There are three types of water in soils depending on the primary

forces that act upon the water at each scale:
1. Hygroscopic water
2. Capillary water
3. Gravitational water
Figure 12: The three types of water in soils, Water can also
be present as vapor in the empty macropores.
(www.tutorvista.com)
The forces present in soil-water matrix are important.

Cohesion is the attraction of the water molecules for one
another. Adhesion is the attraction of the of the water
molecules to the soil particles at the soil-water interface.
Water has dipole characteristics. Each water molecule has a positive charge on one end and a negative charge on
the other. This influences the retention of water in all soils and especially in clayey soils since clay particles are
negatively charged. The water that adheres to soil particles is called hygroscopic water and the bond is relatively
strong for a short distance at the molecular level. Generally, hygroscopic water will not easily transpire if at all. It
is the last of the soil water to remain in the drying process.
The surface tension effects of soil water are more easily observed at the water-air interface. The attraction of the
air molecules for the water molecules is less than the attraction of the water molecules for one another. Thus,
there is a net downward force of the water at the surface. Essentially, this causes the water to behave as through
its surface is a stretched membrane. The surface tension effects causes capillary action in the micropores of the
soil. Capillary action resists gravitational forces.
The larger openings of a soil are usually referred to as macropores. These can depend on the particle size of the
soil, drying cracks, and even tunnels formed by worms and other burrowing animals. The water in macropores is
primarily influenced by gravitational forces.
When and soon after a soil is saturated by a heavy rainfall or irrigation, there is usually some surface water
remaining that hasn’t infiltrated. Below the surface the excess water in the soil macropores begins to percolate
away due to gravity, and the ET process begins (no ET while raining!).
The drying process driven by ET usually continues for about 2 to 3 days until the surface water has evaporated
and the downward movement of soil water below the surface decreases. All the moisture that is left in the soil is
hygroscopic and capillary water. Air, and some water vapor but primarily air, is left in the macropores.
8
D R A F T V5b
The soil (moisture) is said to be at “field capacity” (FC) at this stage. FC is usually expressed as a percentage. The
value is somewhat arbitrary because equilibrium is never really reached and downward movement, although
decreasing significantly, never ceases. Although there have been suggestions to discontinue the use of the term
FC, it is still a useful term and continues to appear in the literature. FC can depend on other factors such a
temperature.
As the drying process continues, a conditional point will be reached when plant roots will no longer be able to
absorb the remaining moisture (hygroscopic) in the soil. When plants can no longer absorb water, they will begin
to wilt (wilting point or WP). If the damage isn’t severe and available soil water is restored, the plant can recover.
When the plants can no longer recover, the “permanent” wilting point (PWP) is said to have been reached. The
permanent wilting point coefficient Cpwp is also usually expressed as a percentage (ratio) of the soil moisture that
is present when the PWP is reached relative to the dry weight of the soil. The PWP can vary with the type of
plant.
Available water for a specific plant is defined as the permanent wilting point subtracted from the field capacity.
AW = FC -PWP (3)
AW is basically capillary water retained by the soil. The retention capacity and drying rate of soils depends on a
number of factors including soil type. Sandy soils dry faster than clayey soils. Soils with a good plant cover dry
slower than bare soils.
Infiltration
Infiltration is a basic process of the hydrologic cycle. The amount of water in a soil is a function of antecedent
moisture remaining from earlier precipitation events (or ground water) and the infiltration rate relative to the
precipitation rate of a current storm event. Runoff occurs when the initial abstractions from surface depressions,
etc., have been satisfied and the rainfall intensity (or irrigation rate) exceeds the infiltration rate of the soil. The
infiltration rate depends on a number of other factors such as soil type and depth of soil and land use cover.
Infiltration occurs at the surface. Once water enters the soil, the movement of the water is termed percolation.
Percolation can contribute to ground water, reappear as quick return flow, or spring flow all of which do not
normally contribute to ET.
Other losses can also occur. Not all precipitation will become direct infiltration. Some will be lost to plant
interception. Some precipitation will be lost to initial abstractions and interception into drying cracks and into
fractured rock. Therefore, the assumption that I = ET is simplistic and not necessarily true. The infiltration rate of
sandy soils is greater than clayey soils. The infiltration rate of compacted soils is less than undisturbed soils. The
infiltration rate of undisturbed soils is less than recently cultivated fields.
Rainfall events can also be classified in terms of runoff as being either non-producing or runoff producing events.
Rainfall events in Tarija are generally light and generate little or no runoff. The soils are typically dry to a considerable
depth. Precipitation that does infiltrate into the soil only wets the soil near the surface and penetrates no deeper. The
shallow wetting of the soils means that the drying process is faster and is usually completed before the next rainfall
event. Soils can return to the dry state in as little as three days although more study is again needed.
There are multiple approaches for estimating losses due to infiltration. Perhaps the most reliable method is the use of
paired data consisting of recorded rainfall depths over a watershed and runoff volume recorded by a streamflow gage at
the watershed outlet. The connecting parameter between depth and volume is the watershed area “A”.
Q = A (P – I) (4)
Where:
Q = Rate if runoff volume in m3/seg, or ft3/sec
A = The watershed area
P = Precipitation depth
I = Infiltration depth (and Ia)
A, P, and I must ultimately be in compatible units of measure.
The streamflow gage only sees the flow without “knowledge” as to how it was produced. If the rainfall amount is
known and assumed to be distributed uniformly throughout the watershed, a reasonable value of infiltration can
be estimated.
9
D R A F T V5b
The method can have problems. For example there can be poorly calibrated streamflow gages, the existence of
underflow that is not detected by the gage, or a bed mobility that exaggerates flow depth. However, the paired
data concept can essentially eliminate the need for infiltration concepts which lack other data.
Adequate streamflow data without paired precipitation can also be used to determine regional water availability.
However, the topic is beyond the scope of these notes.
Infiltration models
The alternative to the paired data concept is to use an infiltration model. Two of the more popular models that are
available are the Green-Ampt (conceptual) and the Natural Resources Conservation Service (NRCS) runoff curve
number (RCN) method (empirical). RCN values typically range from 60 to 100 (100 is totally impervious)
depending on soil type and cover.
The RCN equation is popular, useful, and includes consideration of initial abstraction. It takes the form of:
Q = (P – Ia)2 / [(P – Ia) + S] (5a)
Q = Runoff depth (mm or inches)
P = Rainfall depth (mm or inches)
Ia = Initial abstraction (mm or inches), Ia = 0,2 S
S = Potential maximum retention depth (mm or inches)
S= (1000 – 10 RCN) / RCN US customary (5b)
S= (25 400 – 254 RCN) / RCN System International (5c)
Infiltration methods generally apply to valid for relatively deep (0,5 m+) soils for average precipitation events.
The depth value can be less in arid regions when storm depths are small. If only rock is present or only shallow
soils exist over rock, the soil can become saturated quickly. The problem then becomes ground interception and
percolation and not true infiltration. The issue is the need to determine how much water will enter the rock and
how much will be converted to runoff.
Research and data on rock interception are generally lacking. An initial, conservative approach is to assume that
the rock is impermeable. However, rock can be in a state of decay, contain cracks and fissures, and open faults.
Rock can also include bedding planes (horizontal, inclined, or even vertical) or layers much like the “peels”
separating the successive layers of onions. This can be the case for granite. The uncertainty associated with
percolation through rock can be aggravated by situations where the openings are filled by soil or debris.
Figure 13: The bedding planes of tilted sedimentary rock are exposed on the east side of the Franklin
Mountains near El Paso, Texas. Open bedding planes on the east side tend to intercept surface runoff.
Surface runoff is greater on the west side given the same rainfall event
occurring over both sides of the mountains. (adapted by the author from
staticflickr.com)
Soil conditions in the El Valle Central can be classified into two categories.
Shallow soils over rock in the hills and mountains, and deep lacustine
deposits in the valley. The shallow soils over rock appear to be the dominant
watershed condition.
Figure 14: Shallow soil over rock is the dominant soil conditions at the
higher elevations in El Valle Central watershed. (author)
10
D R A F T V5b
One of the principal problem of applying infiltration concepts for water balance studies is that antecedent soil
moisture will impact the rate of infiltration and thus runoff. AMC moisture conditions are problematic.
The process when considering the change of storage of soil moisture is to sequentially determine the initial
antecedent moisture condition (AMC), then add the infiltration that occurs due to the precipitation event to
determine the new, adjusted AMC. The final step is to subtract the ET amount that occurs between the first and
second rainfall events in order to determine the AMC to be used for the next event. Note that ET is not considered
to occur during a precipitation event.
At one time this was considered a complex process and an excuse for using the simpler ET methods. It still is but
the process and soil moisture now can be estimated through computerized Soil Moisture Accounting (SMA)
algorithms used in continuous simulation modeling. The USACE software package HEC-HMS includes SMA
capability.
Other similar “hydrologic” computerized models available including both SMA and ET features include the
Témez model, HBV model, Sacramento model, and the WEAP model. The list is not all inclusive. The Témez
and WEAP models have been used in Bolivia (Villena).
In Tarija, when soils are wetted during the rainy season, the depth of the wetting front is limited to a shallow depth
due to the typical short, low depth storm durations. There simply is not enough water to penetrate and wet the soil
to a greater depth. There are other incidental causes that can reduce infiltration and increase runoff. These can
include crust formation near the soil surface. Crust formation usually occurs on relatively bare soil in arid regions.
Causes can include repeated wettings and drying of the soil, wind removal of fines (deflation), the presence of
microbes and algae, and cementitious materials present in the soil. A temporary phenomena is the development of
a pressure layer or “aircushion” when rain impacts the surface of a bare soil.
Soils can then rapidly dry out before the next rainfall event (the average interevent period between precipitation
events is five days) in Tarija. Bastian reported similar infiltration/high runoff conditions aggravated by the
appearance of a surface-sealing slime during rainfall events. He also reported the quick drying of soils with the
complete desiccation of the root zone during the dry season.
Overall, the existence of dry conditions in the deeper lacustine soils can greatly simplify the application of
infiltration concepts in determining runoff for Tarija. AMC for infiltration conditions can essentially be
considered to be reset to dry conditions at the beginning of every precipitation event.
The NRCS uses four soils classification/types for their procedures. These and associated characteristics are:
• Soil Type A – These soils such as unconsolidated sands are associated with high losses and low runoff
• Soil Type B – Slightly more impermeable than A soils
• Soil Type C – Slightly less impermeable than D soils
• Soil Type D – These soils such as tight clays are associated with low losses and high runoff
A practical basin-wide RCN can be estimated
assuming Type “C” soil and a “poor condition”
brush-weed land use cover. A runoff curve number
(RCN) value of 77 is about right for these conditions.
The 77 RCN value has been tested twice in
inundation studies conducted by UCB students and
correlates fairly well with historical events. In
addition, a recently completed preliminary soil
infiltration study of the lacustine deposits for the
Lower La Vittoria and the El Monte watersheds also
indicates that soils in the two watersheds closely
resemble “C” type soils. Figure 15: The typical land use or cover conditions in El
Valle Central is a brush-weed mixture in “poor” condition.
ETr will typically be low. A value of about 580 mm is often
cited for Tarija, often without reference to an annual
average rainfall depth. (author)
11
D R A F T V5b
Figure 16: In contrast to the general conditions shown in Figure 15

is the humid micro climate of El Rincon de La Vittoria, an area of
about 33 km2 in the Reserva de Las Samas. Diversions from the
watershed account for about 30 percent of the City of Tarija’s
potable water. (author)
A study of the rocky areas that dominate the El Valle Central watersheds was not conducted by UCB. The
possibility of an investigation is being considered. Again, precipitation does not infiltrate into rock. Tight, intact
rock is practically impermeable. Highly fractured, faulted rock will intercept flow. Rasmussen has suggested RCN
values of 75 to 85 for dry condition, highly fractured rock, and that values can exceed 90 under wet conditions.
Figure 17: Exposed rock in El Valle Central is often highly fractured. (author)
Figure 18: Urban infiltration conditions generally represent an impermeable

surface regardless of soil type due to the high percentage of street and roof cover.
Runoff is high and ET ≈ 0. (hoybolivia)
Initial abstraction “Ia”, the forgotten “orphan” when assuming ET = I
Initial abstraction “Ia” is normally lumped with infiltration and is quickly forgotten. However, it can be important,
especially if the typical precipitation depths for the region are small. This is the case for Tarija.
As an explanation, note that the monthly precipitation depths for both April and October in Example 3 (that
follows later) are both 20 mm. The runoff “Q20” for 20 mm of precipitation is only 0,3 mm.
Since Q20 = P – I,. “I” is supposedly 19,7 mm (0,3 mm = 20 mm – 19,7 mm). However, if the value of Ia (15,2
mm for a RCN of 77) is isolated from “I” only 4,5 mm can be considered “true” infiltration that is available to
replenish soil moisture. The 15,2 mm of Ia is “trapped” at the surface as puddles. Initial abstraction can also
include canopy interception (water trapped on tree and plant leaves).
This puddled water may or may not eventually seep into the ground but much of it will quickly disappear as
surface evaporation. When Ia is associated with low precipitation depth, it cannot be assumed that infiltration =
ETr and that P – ETr = runoff. Without Ia, water available to replenish soil moisture, ETr will need to be adjusted
to a lower value, at least conceptually, as a measure of total loss. Thus, computed runoff will be higher.
12
D R A F T V5b
In regions where rainfall depths are typically higher, the ratio of Ia to true infiltration becomes smaller but never
really losses its significance. Ground interception and Ia are also difficult to separate in rock. Finally, note that
that the Ia and infiltration values for the RCN procedures are empirically-based and not physically-based. RCN
was used for the discussion only as a demonstration but the use of other infiltration models should be considered.
ET
ET is the process of evaporation and plant transpiration. which in combination is usually referred to as
evapotranspiration “ET”. ET removes moisture from the soil. There are many factors that impact ET. However,
the overall driver in the ET process, perhaps in the end and all factors considered, is the vapor pressure gradient
between the soil and the atmosphere.
If the atmosphere is humid, the vapor pressure gradient may have a value that somewhat limits uptake (Δ Vapor
pressure between atmosphere and soil is small and thus so is ET). If the atmosphere is dry, the atmospheric
demand for uptake from the soil will likely be greater and thus so will ET. Again, radiation (sunlight, ground
radiation) and its surrogate temperature, wind, infiltration, and a host of other parameters also have impact on ET
but the vapor pressure gradient is the key.
Potential evapotranspiration “PET” is the maximum evapotranspiration that can occur based on climatic demand
alone. Real (or actual) evapotranspiration is referred to as “ETr”. ETr depends on the amount of moisture that is
actually present in the soil. That is, PET identifies what “can” potentially occur and ETr identifies what “will”
likely occur. Another form, “ETo” is ET for a referenced crop such as a grass (the use of alfalfa is common).
In an extreme case, if the soil is completely dry then there is nothing available to evaporate or transpire. In
general, ETr will not be as high as but can approach PET when adequate soil moisture is present. The high
availability of soil moisture in swampy conditions is a good example.
The Penman-Monteith equation, as presented in FAO 56 (1998), is the standard and the most complete ET
method available in terms of the influencing climatic parameters. Its drawback is that it requires a wide range of
climatic data that are often not available. Some simple ET models such as the Turc method are at the opposite end
of the FAO 56 detailed process. The Turc method only considers precipitation and radiation (as temperature).
These simple methods are useful when climatic data is limited as in the case of many developing nations. The
simple methods are also useful for preliminary evaluations.
Caution must be used when using the simpler methods such as the Turc (1961), Coutagne (1935), and similar
equations. Often they were developed in their original form for a specific purpose, and specific regions with
specific watershed, plant cover, soil types, and climatic conditions. The Turc method has been identified as being
developed for irrigation estimates using data for the more humid, relatively uniform conditions of western Europe
(there is some conflict in the description of the secondary citations). The Coutagne equation was developed using
climatic data for the province of Lyons in France. It appears that neither method was intended for use in semi-arid
regions and regions with distinct wet-dry periods and watershed complexities such as El Valle Central.
There are other relatively simple ET methods that have been tested and are available for use in semi-arid regions
(assuming the same precautions are used as previously described). Some of these include:
• Priestley-Taylor (1972)
• FAO 24, Radiation Method (1977)
• FAO 24, Blaney-Criddle (1977)
• Hargreaves-Samani (1985)
ET is often used for water availability studies when infiltration data are lacking and because the lumped methods
are simply easier to apply (sometimes too easy) . However, infiltration and ET are not the same thing, especially
when a large areal cover of rock or urban development are present in the watershed. Once again, water generally
does not infiltrate into rock and rock does not have a water content available in the same sense as soil. Therefore,
when rock dominates the watershed, the application of ET concepts as a primary model is very questionable.
However, when a watershed includes mostly soils, the soils are generally dry, and the typically rainfall depths are
relatively small and light, ET and infiltration have been considered by many to be essentially equal to one another
and interchangeable. This seriously ignores the issue of initial abstraction and is an unwarranted extrapolation.
13
D R A F T V5b
Additional information on selecting an ET model can be found in Table 2 of Fisher, et al:
Figure 19: Table 2 from Fisher, et al

Interevent Precipitation
Several desirable pieces of climatic data are usually missing or at least have to be derived from weather data.
Most “humid” regions of the world have frequent precipitation events throughout the year. Both Tarija and Texas
have a dry and rainy season. Fortunately for Tarija, the rainy season occurs in summer in the southern hemisphere
and is coincident with the growing season when water demand for plants is at its greatest. This is not the case for
Texas when the dry period (summer in the northern hemisphere) coincides with the growing season. More
irrigation is required for the growing season in Texas.
The interevent time between rainfall events in Tarija during the wet period is about five days. This corresponds
with the former practice (1964 to 1993) by the NRCS to apply a five-day antecedent precipitation to determine
AMC. By 1993 it was determined that antecedent precipitation had little impact on AMC. The term used now is
antecedent runoff condition (ARC). The use of AMC persists in the literature.
14
D R A F T V5b
Regardless of terms, the norm can be dry, average, or wet. West Texas is dry, east is wet, and central Texas
somewhere in-between.
Whether rainfall is available throughout the year or is seasonal can also have a significant impact on the
availability-demand relationship for a region and the method of evaluation.
Demand, safe yield, and allocation concepts
A complete discourse of the topics in this section are beyond the scope of this document but important for water
supply studies. They will only be touched upon briefly.
Demand. Demand consists of identifying the uses and each associated amount of the users in the watershed..
Generally, agriculture accounts for about 90 percent of the potential use and municipal /domestic
consumption are include in the other 10 percent. Highly industrialized areas can change the percentages.
Agricultural demand can vary depending on the growing season and is highly variable. Municipal and
industrial use are usually nearly constant. Demand projections are used for planning.
Often demand can exceed the regional supply and water must be imported from other areas. The water
demand-supply relationship of Los Angeles in the state of California in the United States is a classic
example. Los Angeles is located in a desert. It never was capable of supporting its current population.
Supply. Supply is the principal topic of this document.
Safe yield (reliability of meeting demand). Safe yield is a comparison of demand with the concept of a
relatively guaranteed supply for a specific period. Historic records are used to determine the probability
that a specific amount of water can be met. A common minimum criteria for agriculture is 80 percent and
can be as high as 95 percent for important commercial crops.
For example, an 80 percent annual probability means that the demand needs can likely be met 4 out of 5
years and that there will be deficit 1 out if 5 years. Storage reservoirs and ground water may be needed to
offset the deficit period. However, climate change trends may alter the supply relationship. An
unexpected shift in demographics may also alter the use projections. Uncontrolled contamination and
natural catastrophes may reduce the usable supply.
Allocation and the legal right to use a specific quantity. Often overlooked, allocation of water within a legal
system to fairly distribute specific quantities to users should be considered. The point is to allocate, the
government should be considered the ultimate owner of the available surface and ground water supplies.
Existing, first users in time are usually granted priority. An adjudication process may be needed to sort
out the legalities if a system does not exist. The laws should recognize the priority of municipal use
during droughts.
Even the concept of a just allocation system can be complicated by natural conditions. A typical
characteristic of the watercourses and streams of El Valle Central is that they often have a high bed
mobility. The deep alluvium of the channels can absorb surface flow and transform it to underflow (see
Figure 5).
This is very obvious in the La Vittoria watershed. The watershed is divided roughly in two parts
consisting of a narrow canyon in the upper half and a broader valley in the lower area. Streamflow is
often present in the upper canyon but disappears from the surface within 600 +/- m after exiting the
canyon and entering the broader valley. The water is still there but it exists as underflow. Infiltration
galleries or wells are needed to recover the water for the users in the lower part of the valley.
Advances in the art
This section will briefly touch on advancement in the art of ET and remote sensing. An example is the METRIC
(Mapping EvapoTranspiration at high Resolution with Internalized Calibration) computer model. METRIC,
which was developed by the University of Idaho, computes and maps ET using Landsat satellite data. The
Landsat satellite system collects thermal data and is preferred since it is only operational satellite system that has
a data resolution sufficient to capture data needed for irrigation applications at the scale of individual agricultural
fields. Remote sensing applications that can be used for regional water balance applications also exist.
15
D R A F T V5b
Figure 20: Remote sensing concepts (CRISP)
Figure 21: Remote sensing capture of sensible and

latent heat fluxes over rice paddies (Liou)
Figure 22: Table 1 Remote sensing models from Liou
16
D R A F T V5b
SUMMARY of PART 3
Whether for water availability studies or irrigation budget applications, it has been demonstrated in these notes
that water balance studies can be complicated.
Simplified (lumped) and complete ET approaches exist for irrigation applications on the scale of cultivated fields.
Their primary advantage is that they reflect climatic conditions. The primary disadvantage is that they seldom
reflect watershed conditions. However, they have been applied and reportedly have provided reasonable estimates
for watershed yields at a much larger scale when data was scarce.
Infiltration models can be used for regional water availability studies but are seldom used for irrigation
applications. Their primary advantage is that they reflect watershed conditions. Their disadvantage is that they do
not reflect very few climatic conditions
Both ET and infiltration are specific elements of the hydrologic cycle. How the two are used depends on the
intended purpose of a study. Engineering experience and judgment are essential for the proper selection and
application of a model.
Advances in technology have improved both approaches. Advances in modeling techniques for continuous
simulation offer the opportunity to use both in the same model in accordance with their respective roles in the
hydrologic cycle. That is, apply a infiltration module during the precipitation events and an ET module for the
time span between rainfall events.
Water balance methods, whether for watershed availability applications or irrigation budgets, cannot be
approached mechanically. Modelers must examine their data, understand the full range of methods that
are potentially available, and know if their regions are compatible with the methods.
Part 4: Simple Examples

Lumped ET methods
The Turc (1961) method is a popular lumped method developed for the more humid regions of western Europe.
Vegetation can be more lush and some regions can be heavily cultivated. There are variations of the equation for
considering time scale (a variation is available for an incremental time period of 10 days) and regional humidity. The
break point for humidity is 50 percent. The average humidity for Tarija is 59 percent. The adjustment applies to regions
with less than a 50 percent average. The Turc equation for annual ETr for conditions of (P/L > 0,316, otherwise ETR =
P) takes the form of:
(6)
Where:
ETr = Actual evapotranspiration (mm / year)
P = Annual precipitation (mm / year)
L = A regression value for temperature
L = A+25(T) + 0,05 (T)n, “A” is Turc’s constant = 300, n = 3. Turc considered “n” = 3 a best fit but
apparently there was some thought by Turc that “n” could also be 2, He originally retained 2 for
computational simplicity. Lebecherel, p 7509
T = Average annual temperature.
Example 1
A consultant study for a water availability analysis of a watershed was recently completed.
The average annual temperature for the total watershed is given as 14,7o C. The average annual precipitation is
given as 721,9 mm. Determine ETr, the runoff, and ratio of runoff to precipitation in terms of depth. It is not clear
how many years of record were referenced by the consultant nor the specific years. Use Turc’s equation to check
the consultant’s results.
17
D R A F T V5b
Assumptions: Q = P – ETr, and ETr = infiltration. P and T are the only distinguishing regional factors.
ETr
First calculate “L”.
L = 300 + 25 (14,7) + 0,05 (14,7)3, L = 826,3, (if n = 2, L = 678)
Check P/L
P/L = 721,9 / 826,3, = 0.874 > 0,316 OK
Calculate ETr
ETr = 721,9 / {0,9 + [(721,9)2 / (826,3)2]}1/2 , = 559,8 mm, (if n = 2, ETr = 506,2 mm)
The value checks fairly well with the consultant’s value of 531.6 mm and that of other citations for
studies in the region.
Calculate runoff
Q = P -ETr, = 721,9 mm – 559,8 mm, = 162,1 mm of runoff, (if n = 2, Q = 215,7 mm)
Calculate % Q / P
% = 162,1 mm / 721,9 mm, = 22 % of the precipitation is transformed to runoff, (if n = 2, Q = 30 %)
Example 2
Data is provided in Table 1 (metric and US customary) for what is assumed to be the Upper Guadalquivir River
above the Angosto San Luis. The average annual temperature is given as 19 o C. The average annual precipitation
is given as 490 mm. Determine the metric values of ETr, the runoff, and ratio of runoff to precipitation in terms of
depth.
ETr
First calculate “L”.
L = 300 + 25 (19) + 0,05 (19)3 , L = 1117,9, (if “n” = 2, L = 1136, there will be very little difference)
Check P/L
P/L = 490 / 1117,9 = 0,438 > 0,316 OK, but we are approaching the point where ETr ≈ P
Calculate ETr
ETr = 490 / {0,9 + [(490)2 / (1117,9)2]}1/2 , = 468,9 mm
Calculate runoff
Q = P -ETr, = 490 mm – 468,9 mm, = 21,1 mm of runoff
Calculate % Q / P
% = 21,1 mm / 490 mm, = 4,3 % of the precipitation is transformed to runoff
The data looks suspect and nonrepresentative, it is probably due to the temperature value which appears high.
However, not only is there a mix of years of record, the time spans may be different and offset. Furthermore, the
Upper Guadalquivir includes the Tolomosa Grande (San Jacinto) which is one of the wettest in El Valle Central.
For the consultant’s average temperature of 14,7o, ETr would be 391,5 mm. Runoff would be 98,5 mm indicating
that about 20 % of precipitation is transformed to runoff. Again, the percentage of runoff agrees well with the
Example 1.
Example 3
As a demonstration of infiltration methods, use the data in Table 1 to evaluate runoff in metric and US Customary
values. Use the NRCS RCN method to evaluate runoff for different time increments.
18
D R A F T V5b
Table 1: Climatic Data for Examples 2 and 3 Source: Weatherbase
http://www.weatherbase.com/weather/weather.php3?s=46358&refer=wikipedia
Average Precipitation Years on Record: 15
ANNUAL JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
mm 490 120 80 60 20 --- --- --- --- 10 20 40 90
Average Precipitation Years on Record: 15
in. 19.3 5 3.5 2.4 1 0.1 --- 0.1 0.1 0.6 0.8 1.9 3.8
Average Number of Rainy Days Years on Record: 15
Days 35.3 7.7 5.8 4.1 2.2 1 --- 0.5 0.5 1.6 2 3.7 6.2
Assumptions: Q = P – I - Ia. Soil type and land use cover are important.
Compute preliminary data:
P avg = 490 mm / 35 days, = 14,0 mm (0,28 inches) for each daily event
Interevent days = [(7 months wet / 12 months per year) (365 days per year) – 35 days wet] / (35 days wet)
Interevent days = 5 days
Average Precipitation Depths (monthly and daily averages)
P jan = 120 mm / 8 days, = 15 mm (0.60 inches)
P feb = 80 mm / 6 days, = 13 mm (0.50 inches)
P mar = 60 mm / 4 days, = 15 mm (0.60 inches)
P abr = 20 mm / 2 days, = 10 mm
P may = 0
P jun = 0
P jul = 0
P aug = 0
P sep = 10 mm / 2 days, = 5 mm (0,2 inches)
P oct = 20 mm / 2 days, = 10 mm (0.4 inches)
P nov = 40 mm / 4 days, = 10 mm (0.4 inches)
P dec = 90 mm / 6 days, = 15 mm (0.6 inches)
The monthly total values provide a better description of the variation in precipitation depth than the average of the
total annual value. We can also compute monthly interevent periods between rainfall events. For example, the
interevent time for January would be about 4 days which isn’t much different than the annual value of 5 days.
Specific daily values are not available in the Table and so the daily averages based on the monthly depth and
number of rainy days is the best we can do.
Runoff based on RCN method
Recall that runoff using the NRCS RCN method can be computed as follows:
Q = (P – Ia)2 / [(P – Ia) + S] (5a)
19
D R A F T V5b
Where:
Q = Runoff depth (mm or inches)
P = Rainfall depth (mm or inches)
S = Potential maximum retention depth (mm or inches)
S = (1000 – 10 RCN) / RCN US customary (5b)
S = (25 400 – 254 RCN) / RCN System International (5c)
Ia = Initial abstraction (mm or inches), and
Ia = 0,20 S (Mockus, original and used in this presentation)
Note: Ia = 0.05 S (per research by Fennesey and Hawkins)
Assume the soil type representative of the watershed is “C”.
Assume that the predominant cover is a brush-weed mix in poor condition and so RCN =77. Therefore:
S = (25 400 – 254 (77) / 77, = 75,9 mm (2.99 inches)
Ia = 0,2 S, = 15,2 mm (0.6 inches)
The term “Ia” is an important value. Runoff cannot occur until the initial abstractions are satisfied. Recall that the
average daily precipitation value is 14 mm based on annual averages. The average daily value based on monthly
averages is 15 mm. Very little runoff, if any, will occur considering these values relative to an Ia of 15,2 mm.
Scale is an important concept and the use of “averages” can be deceptive.
Annual runoff based on average annual rainfall

Q = (490 – 15,2)2 / [(490 – 15,2) + 75,9], = 409 mm (16 inches) of runoff
Calculate % Q / P
% = 409 mm / 490 mm, = 83,5 % of the precipitation is transformed to runoff
This is NOT consistent with experience. The distortion is due to the limiting value of S = 75,9 mm for
infiltration relative to the depth of precipitation, and time scale. “S” limits the maximum amount of
infiltration that can occur during an event of selected duration (1 year in this case).
The method used by the modeler must be completely understood !
Accumulated annual runoff based on average monthly rainfall
However, many individual, shorter events can occur over a year. They can also be separated by interevent times
that allow for some drying. The interevent time in Tarija is about five days which allows for drying of soil
moisture and adjustment, if not a complete reset, for the RCN.
Our data indicates:
P jan = 120 mm monthly. 120 / 8 days, = 15 mm (0.60 inches) daily average
P feb = 80 mm monthly. 80 / 6 days, = 13 mm (0.50 inches) daily average
P mar = 60 mm monthly. 60 / 4 days, = 15 mm (0.60 inches) daily average
P abr = 20 mm monthly. 20 / 2 days, = 10 mm (0.40 inches) daily average
P may = 0
P jun = 0
P jul = 0
P aug = 0
P sep = 10 mm monthly. 10 / 2 days, = 5 mm (0.2 inches) daily average
P oct = 20 mm monthly. 20 / 2 days, = 10 mm (0.4 inches) daily average
P nov = 40 mm monthly. 40 / 4 days, = 10 mm (0.4 inches) daily average
P dec = 90 mm monthly. 90 / 6 days, = 15 mm (0.6 inches) daily average
20
D R A F T V5b
The daily average based on monthly values presents a problem. Specific daily values are highly desirable and
only averages are available. We can only proceed with the evaluation of the infiltration method using monthly
average values.
In order of magnitude the monthly precipitation depths are:
120 mm, 90 mm, 80 mm, 60 mm, 40 mm , 2 x 20 mm, 2 x 10 mm, and 3 x 0 mm.
Precipitation depths of less than 15,2 mm do not meet the threshold for producing runoff.
Q 120 mm = (120 – 15,2)2 / [(120 – 15,2) + 75,9], = 60,8 mm
Q 90 mm = (90 – 15,2)2 / [(90 – 15,2) + 75,9], = 37,1 mm
Q 80 mm = (80 – 15,2)2 / [(80 – 15,2) + 75,9], = 29,8 mm
Q 60 mm = (60 – 15,2)2 / [(60 – 15,2) + 75,9], = 16,6 mm
Q 40 mm = (40 – 15,2)2 / [(40 – 15,2) + 75,9], = 6,1 mm
Q 20 mm = (20 – 15,2)2 / [(20 – 15,2) + 75,9], = 0.3 mm, = 0,3 x 2, = 0,6 mm
Q 10 mm = (10 – 15,2)2 / [(10 – 15,2) + 75,9], = 0 mm,
accumulated Q total = 151,6 mm compared to 409 mm for the annual “distorted” value.
% = 151 mm / 490 mm, = 30,8 % of the precipitation is transformed to runoff
Possible improvements: Consider variation in soil types, presence of rock, variation in cover.
Initial abstraction “Ia”, the new “demon” issue introduced in Example 3

Initial abstraction is normally lumped with infiltration and is quickly forgotten as a possible issue. However, it can
be important and especially if the typical precipitation depths for the region are small. This is the case for Tarija.
As an explanation, note again that the monthly precipitation depths for both April and October in Example 3 are
both 20 mm. The runoff “Q20” for 20 mm of precipitation is only 0.3 mm. Since Q20 = P – I, “I” is supposedly
19,7 mm (0,3 mm = 20 mm – 19,7 mm). However, if the value of Ia (15,2 mm for a RCN of 77) is isolated from
“I” only 4,5 mm is “true” infiltration that is available to replenish soil moisture. The 15,2 mm of Ia is “trapped” at
the surface in puddles. It may or may not eventually seep into the ground but much of it will disappear as surface
evaporation. In such cases (low precipitation depth), it cannot be assumed that infiltration = ETr and that P – ETr
= runoff. Without the volume of water trapped by Ia available to replenish soil moisture, ETr will need to be
adjusted to a lower value at least conceptually. Runoff will be higher than if ETr is only computed with a simple
method. In regions where rainfall depths are typically higher, the ratio of Ia to true infiltration becomes smaller
but never really loses its significance.
Note that that the Ia and infiltration values for the RCN
procedures are empirically-based and not
physically-based. RCN was used for the discussion only
as a demonstration.
Figure 23: Climatic conditions (circle) for Tarija and

typical biomes which include woods, scrublands,
grasslands and desert. (Riccardo Rigon)
21
D R A F T V5b
Example 4
The results of Example 3 indicated that the Ia and infiltration combination for an RCN of 77 would result in less
soil moisture availability and higher runoff than that determined by Turc’s equation. The NRCS ECN method has
been available and tested for over 60 years. Although empirical and not theoretically-based, it has proven to
provide fairly good results representative of a watershed’s soil, land use cover, and antecedent moisture
conditions. In Problem 4, we will use “reverse engineering” to extract and possible cover condition that would be
representative of the Turc results of Example 1 and Example 2. We will continue to assume soil type “C”.
The RCN value of 77 was developed assuming a brush-weed cover in poor condition which can be fairly
representative of an arid climate similar to the general conditions of Tarija.
Consider change of condition of land use cover.
The condition of the assumed, general brush-weed cover was considered to be poor, RCN = 77. Assume a “good”
condition (denser, healthier vegetation) and recalculate.
RCN = 65, S = 136,6 mm (5.38 inches), and Ia = 27,4 mm (1.08 inches).
Compute the runoff depths
Q 120 mm = (120 – 27,4)2 / [(120 – 27,4) + 136,6], = 37,4 mm
Q 90 mm = (90 – 27,4)2 / [(90 – 27,4) + 136,6], = 19,7 mm
Q 80 mm = (80 – 27,4)2 / [(80 – 27,4) + 136,6], = 14,6 mm
Q 60 mm = (60 – 27,4)2 / [(60 – 27,4) + 136,6], = 6,3 mm
Q 40 mm = (40 – 27,4)2 / [(40 – 27,4) + 136,6], = 1,1 mm
Q 20 mm = 0, Q 10 mm = 0
Accumulated Q total = 79,1 mm
This is closer to the ETr values of Example 1 and modified Example 2. The results indicate that the general
vegetative cover for El Valle Central would have to be fairly lush for the Turc method to be valid and this is
not the case.
Using similar methods for comparison and validation
The consultant used the Coutagne method (1935) as partial validation of the results of Turc method. The
Coutagne method was developed for the Lyon region of France. General climatic data for the Lyons region is
provided as follows:
Figure 24: Climatic data for Lyon France. (Wikipedia at: https://en.wikipedia.org/wiki/Lyon )
The readers are left to make their own conclusions as to the applicability of both of the ET (Turc and Coutagne)
methods for Tarija. You may want to take a look at Figure 17. You may also want to calculate ETr.
22
D R A F T V5b
Conclusions based on the Examples 1 through 4
Compare the ETr values of Example 3 to Example 1 and the “modified” Example 2. The Example 1 and the
modified Example 2 values are similar (≈ 20 to 22 percent of the precipitation is runoff). Not surprising because
the method in both cases is the same and assumes that the conditions in western Europe that were used to develop
the Turc equation are applicable to the watershed, climatic, and hydrologic processes to those in Tarija
Specific daily values are really needed and better data for the areas that are dominated by rock before any real
conclusions can be made concerning the infiltration method. A better evaluation of the rock may indicate more
runoff. A better understanding of Ia as a factor in replenishing soil moisture is needed.
However, as a “preliminary” study, it appears that runoff for the El Valle Central watershed can range from about
20 to 30 percent of average annual precipitation. The lower runoff values of the Turc method are more
conservative for water supply planning purposes but may limit the potential for future development.
The urgency of having reliable hydrologic models for Tarija is motivation for government support and
pursuit of better and more completed data collection systems.
References used for these notes
Allen, Richard G., Pereira, Luis S., Raes. Dirk, & Smith, Martin (1998). Crop evapotranspiration - Guidelines
for Computing Crop Water Requirements - FAO Irrigation and Drainage Paper 56. Food and
Agriculture Organization of the United Nations, Rome
ASCE Standardization of Reference Evapotranspiration Task Committee (2002). Draft Report,
Standardization of Reference Evapotranspiration. Environmental and Water Resources Institute of the
American Society of Civil Engineers
Bastian, E., & Gräfe, W. (1989). Afforestation with ‘multipurpose trees’ in ‘media lunas’, a case study from
the Tarija basin, Bolivia. Agroforestry systems, 9(2), 93-126
Clarke, I. D., Roose, K., & Riemer, W. (1991). Protection of the Pleistocene Sediments along the rim of the
San Jacinto reservoir in Bolivia. Geological Society, London, Engineering Geology Special
Publications, 7(1), 611-617
Coltori, M., Abbazzi, L., Ferretti, M. P., Iacumin, P., Paredes Rios, F., Pellegrini, M., Pieruccini, P., Rustioni,
M., Tito, G., Rook, L. (2006). The last Glacial mammals in South America: a new scenario form the
Tarija Basin (Bolivia). Springer-Verlag (published online)
Coutagne, Aimé (1935, French). Le déficit d'écoulement et l'évaporation dans la région lyonnaise. Les Études
rhodaniennes, vol. 11, n°3, pp. 359-366
Faure, Paméla F., Coussot, Philippe (2010). Drying of a model soil. Physical Review E : Statistical,
Nonlinear, and Soft Matter Physics, American Physical Society, 82 (3), pp.036303. 10.1103/Phys-
RevE.82.036303
Fennessey, Lawrence A.J., & Hawkins, Richard H. (2012?), The NRCS Curve Number, a New Look at an
Old Tool, https://www.villanova.edu/content/dam/villanova/engineering/vcase/sym-
presentations/2001/B32.pdf
Fisher, Joshua B., Whittaker, Robert J., & Malhi, Yadvinder (2011). ET come home: potential
evapotranspiration in geographical ecology. 8Global Ecology and Biogeography, 20, 1–18
Gafurov, Abror (2010). Water Balance Modeling Using Remote Sensing Information - Focus on Central Asia.
Inst. für Wasserbau
Gruff, R. W., & Thompson, T. H. (1967). A Comparison of Methods of Estimating Potential
Evapotranspiration from Climatological Data in Arid and Subhumid Environments. Contributions to
the Hydrology of the United States Geological Survey, Water-Supply Paper 1H9-M
INE. (2018). Proyecciones de Población Por Sexo, 2017-2020. Bolivia: Instituto Nacional de Estadística.
Retrieved from https://www.ine.gob.bo/index.php/principales-indicadores/item/640-poblacion-del-
municipio-de-tarija-llega-alrededor-de-247-mil-habitantes
23
D R A F T V5b
Gao, Fei, Feng, Gary, Ouyang, Ying, Wang, Huixiao, Fisher, Daniel, Adeli, Ardeshir & Jenkins, Johnie
(2017). Evaluation of Reference Evapotranspiration Methods in Arid, Semiarid, and Humid Regions.
Journal of the American Water Resources Association, Vol. 53, No. 4
GEOBOL (1991). Hoja Geológica Tarija 6629
Hanson, R.L., (1991). Evapotranspiration and Droughts. U.S. Geological Survey
Hawkins, Richard H., Jiang, Ruiyun, Woodward, Donald E., Hjelmfelt, Jr., Allen T., VanMullem, J.E, (2003),
Runoff Curve Number Method: Examination of the Initial Abstraction Ratio.
https://www.researchgate.net/publication/313007483_Runoff_curve_number_method_Examination_
of_the_initial_abstraction_ratio
Lebecherel, Laure, Andreassian, Vazken, & Perrin, Charles (2013). On regionalizing the Turc-Mezentsev
water balance formula. Water Resources Research, vol. 49, 7508–7517
Leib, David I., & Stiles, Thomas C. (undated). Yield Estimates for Surface-water Sources. Kansas Water
Office, Topeka, Kansas, Bulletin 239 – Chapter 6, 158 - 169
Liou, Yuei-An, & Kumar Kar, Sanjib (2014). Evapotranspiration Estimation with Remote Sensing and
Various Surface Energy Balance Algorithms—A Review. Energies
Martin, J.A., & Gray, D.M. (undated). Estimating Evapotranspiration in a Regional Scale. Division of
Hydrology, University of Saskachewan, Saskatoon, Saskachewan
McShane, Ryan R., Driscoll, Katelyn P., & Sando, Roy (2017). A Review of Surface Energy Balance Models
for Estimating Actual Evapotranspiration with Remote Sensing at High Spatiotemporal Resolution
over Large Extents. USGS Scientific Investigations Report 2017–5087
Mockus, Victor (1965). National Engineering Handbook, Section 4, Hydrology, Chapter 21 – Design
Hydrographs. Soil Conservation Service
NRCS (2004). National Engineering Handbook, Part 630 Hydrology, Chapter 10 – Estimating Direct Runoff
from Precipitation
NRCS (2000). National Engineering Handbook, Part 630 Hydrology, Chapter 20 - Watershed Yield
Peschke, G., Dunger, V., & Gurtz, J. (1986). Changes in soil moisture by infiltration and evapotranspiration.
Conjunctive Water Use (Proceedings of the Budapest Symposium, IAHS, Publ.no. 156,
Rasmussen, T. C., & Evans, D. (1963). Water Infiltration into Exposed Fractured Rock Surfaces. Soil Science
Society of America Journal, Vol 57, 324 - 329
Sellers, William D. (1963). Potential Evaporation in Arid Regions. Journal of Applied Meteorology, Vol 3,
98-104
Villena-Martínez, E. M., Torregrosa-López, J. I., & Lo-Iacono-Ferreira, V. G. (2019). DRAFT - State of the
art of hydrological models and its application in the water balance in basins with limited information
Historical references
Turc, L. (1961, French). Evaluation de Besoins en Eau d’Irrigation, ET Potentielle. Ann. Agron., Vol. 12, pp.
13-49.
Turc, L, 1961. Estimation of irrigation water requirements, potential evapotranspiration: a simple climatic
formula evolved up to date. Annals of Agronomy, 12(1): 13–49.
Appendix A
ETr for Lyons, France
Example 5
Use the climatic data for Lyons, France and Turc’s equation to estimate ETr for the region. The average annual
temperature for the Lyons is given as 12,5o C. The average annual precipitation is given as 831,9 mm. Determine
ETr, the runoff, and ratio of runoff to precipitation in terms of depth.
24
D R A F T V5b
ETr
First calculate “L”
L = 300 + 25 (12,5) + 0,05 (12,5)3 , L = 710,2
Check P/L
P/L = 831,9 / 710,3 = 1,17 > 0,316 OK
Calculate ETr
ETr = 831,9 / {0,9 + [(831,9)2 / (710,2)2]}1/2 , = 551,9 mm
Calculate runoff
Q = P - ETr, = 831,9 mm – 551,9 mm, = 280 mm of runoff
Calculate % Q / P
% = 280 mm / 831,9 mm, = 33,7 % of the precipitation is transformed to runoff
Examine the results of Examples 1, 2, and 5. Note the impact of temperature on the results.
Appendix B
Variation of antecedent moisture conditions (AMC) and RCN for part of the year
Example 6
The RCN of 77 developed for The Upper Guadalquivir River watershed is actually for dry (AMC I) conditions
which can be fairly representative throughout the year. AMC I conditions are typical in regions having 600 mm to
700 mm (24 inches to 28 inches) or less average annual rainfall. Regions with average annual rainfalls of about
1000 mm +/- (40 inches +/-) are considered to be representative of AMC II (average) conditions. AMC III (wet
conditions) are the norm when the average annual rainfall exceeds 1500 mm (60 inches +).
These are not absolute values since they assume that the precipitation is distributed evenly throughout the year.
Antecedent moisture conditions can vary considerably in regions having wet and dry seasons.
January (120 mm), February (80 mm), and December (90 mm) are normally the wettest months for Tarija.
Assume that AMC II conditions can also representative for these months for this demonstration. The AMC II
value for an AMC I RCN of 77 is about 89. Recalculate S and Ia for a RCN of 89 and then recalculate the runoff
values for December, January, and February.
For a RCN of 89, S = 31,5 mm (1.24 inches), and Ia = 6,3 mm (0.25 inches)
Compute the new runoff depths and percent runoff
Q 120 mm = (120 – 6,3)2 / [(120 – 6,3) + 31,5], = 89 mm
Q 90 mm = (90 – 6,3)2 / [(90 – 6,3) + 31,5], = 60,8 mm
Q 80 mm = (80 – 6,3)2 / [(80 – 6,3) + 31,5], = 51,6 mm
The depths for all other months are the same as in Example 3 for RCN = 77, and = 23,4 mm
The runoff is now about double of that determined by the Turc method. Of course this is just a demonstration and
a SMA in a continuous simulation model could yield more realistic results.
Appendix C
Consideration of Rock
Example 7
Assume that 55 percent of El Valle watershed is fractured rock. Use the upper-range Rasmussen’s estimate of a
RCN for dry conditions (RCN = 85).
25
D R A F T V5b
Compute a weighted RCN
Table 2: Dry, poor cover RCN Weighted by area
The weighted RCN of 82 for dry conditions (AMC I) and poor cover can then be used to compute runoff as
previously described. Maximum retention “S” = 55,9 mm (2.24 inches), Ia = 11,2 mm (0.44 inches).
The accumulated runoff for Tarija with consideration of rock could now be computed as previously demonstrated
and the results would be higher than that for RCN 77. However, we have accumulated a considerable amount of
interesting data that can be used for estimating monthly and accumulated annual runoff without a more
sophisticated hydrologic model with SMA capability. We will pursue a little more information.
Appendix D
A more complete infiltration model
Example 8
The RCN for the predominant brush-weed cover was considered to be in poor condition for dry antecedent
moisture condition. During the rainy season, the condition of the cover will improve but it will never really reach
“Good”. A “Good” condition is a more dense cover such as that in Zone 2 to the east of El Valle Central. A “Fair”
condition brush-weed cover has a RCN of about 70 for dry conditions, and 85 for the wet season. The weighted
RCN for Fair cover and average AMC II conditions with rock is:
Table 3: Wet, fair cover RCN weighted by area
For a RCN of 90, S = 28,2 mm (1.1 inches), and Ia = 5,6 mm (0.22 inches)
Things have changed somewhat with the refinement process. During the dry season, the threshold rainfall depth
for producing runoff has dropped from 15,2 mm to 11,2 mm. The threshold rainfall depth for producing rainfall
for the wet season, is 5,6 mm.
In order of magnitude, the monthly precipitation depths are:
120 mm, 90 mm, 80 mm, 60 mm, 40 mm , 2 x 20 mm, 2 x 10 mm, and 3 x 0 mm.
Wet season, RCN = 90:
Q 120 mm = (120 – 5,6)2 / [(120 – 5,6) + 28,2], = 91,8 mm
Q 90 mm = (90 – 5,6)2 / [(90 – 5,6) + 28,2], = 63,3 mm
Q 80 mm = (80 – 5,6)2 / [(80 – 5,6) + 28,2], = 54,0 mm
Dry season, RCN = 82:
Q 60 mm = (60 – 11,2)2 / [(60 – 11,2) + 55,9], = 22,7 mm
Q 40 mm = (40 – 11,2)2 / [(40 – 11,2) + 55,9], = 9,8 mm
Q 20 mm = (20 – 11,2)2 / [(20 – 11,2) + 55,9], = 1,2, = 1,2 x 2, = 2,4 mm
Q 10 mm = (10 – 11,2)2 / [(10 – 11,2) + 55,9], = 0 mm, < threshold
Q 0 mm = for 5 months of no rainfall no runoff occurs
26
D R A F T V5b
The results for the more detailed analysis agree quite well with those of Example 6 of Appendix B. That is, when
considering variation in moisture conditions for part of the year but no variation in the condition of the cover or
presence of rock. The runoff reducing effect of an improved cover helps to offset the issues of antecedent
moisture conditions and rock.
Appendix D
A rough estimate of monthly ETr and accumulated ETr using Turc
Example 9
We can return to our monthly precipitation data and add temperature. Recall that these were rather high at
(19 o C vs 14,7 o C for Example 1). Generally, the lower monthly temperature occur in the cooler months when
there is no precipitation. Turc’s equation is also intended for annual ETr. We will play a game and assume that the
precipitation values for each individual month occur every month for a hypothetical year. The individual monthly
precipitation value will thus be multiplied by 12 to obtain a “pseudo” annual value for the equation. We will then
compute the annual ETr value for the specific month. Finally, the pseudo annual ETr values for each month will
be divided by 12 to get a “representative” monthly ETr value and accumulate all for the year. It is a distortion and
we are going to have to use engineering judgment.
P jan = 120 mm monthly, 20 o C
P = 12 (120 mm) = 1440 mm
Calculate “L”
L = 300 + 25 (20) + 0,05 (20)3 , L = 1200, P/L = 1440 / 1200 = 1,2 > 0,316 OK, (n =2, L = 820)
Calculate ETr
ETr = 1440 / {0,9 + [(1440)2 / (1200)2]}1/2 , = 941,3 mm, (n = 2, ETr = 721,4 mm)
Monthly average = 941,3 mm / 12, = 78,4 mm, (n = 2, 60,1 mm)
P feb = 80 mm monthly, 23 o C
P = 12 (80 mm) = 960 mm
Calculate “L”
L = 300 + 25 (23) + 0,05 (23)3, L = 1483,4, P/L = 960/1483,4 = 0,65 > 0,316 OK, (n =2, L = 901,4)
Calculate ETr
ETr = 960 / {0,9 + [(960)2 / (1483,4)2]}1/2 , = 835,9 mm, (n = 2, ETr = 673,1 mm)
P mar = 60 mm monthly, 20 o C
P = 12 (60 mm) = 720 mm
Calculate “L”
L = 300 + 25 (20) + 0,05 (20)3 , L = 1200, P/L = 720 / 1200 = 0,60 > 0,316 OK, (n = 2, L = 820)
Calculate ETr
ETr = 720 / {0,9 + [(720)2 / (1200)2]}1/2 , = 641,4 mm, (n = 2, ETr = 557 mm)
P abr = 20 mm monthly, 20 o C
P = 12 (20 mm) = 240 mm
27
D R A F T V5b
Calculate “L”
L = 300 + 25 (20) + 0,05 (20)3 , L = 1200, P/L = 240 / 1200 = 0,20 < 0,316
Assume ETr = P, = 240 mm
Monthly average = 240 mm / 12, = 20 mm, (n = 2, 20 mm)
P may = 0
P jun = 0
P jul = 0
P aug = 0
P sep = 10 mm monthly, 19 o C
P = 12 (10 mm) = 120 mm
Calculate “L”
L = 300 + 25 (19) + 0,05 (19)3 , L = 1118, P/L = 120 / 1118 = 0,11 < 0,316. (if n = 2, L =793,1)
Assume ETr = P, = 120 mm
P oct = 20 mm monthly, 20 o C
P nov = 40 mm monthly, 20 o C
P = 12 (40 mm) = 480 mm
Calculate “L”
L = 300 + 25 (20) + 0,05 (20)3 , L = 1200, P/L = 480 / 1200 = 0,40 > 0,316 OK, (n = 2, L = 820)
Calculate ETr
ETr = 480 / {0,9 + [(480)2 / (1200)2]}1/2 , = 466,2 mm, (n = 2, ETr = 430.6 mm)
Monthly average = 466,2 mm / 12, = 38,8 mm. (n =2, 35.9 mm)
P dec = 90 mm monthly, 20 o C
P = 12 (90 mm) = 1080 mm
Calculate “L”
L = 300 + 25 (20) + 0,05 (20)3 , L = 1200, P/L = 1080 / 1200 = 0,90 > 0,316 OK, (n = 2, L = 820)
Calculate ETr
ETr = 1080 / {0,9 + [(1080)2 / (1200)2]}1/2 , = 825,9 mm, (n =2. ETr = 665,4 mm)
Monthly average = 825,9 mm / 12, = 68.8 mm, (n =2, 55,4 mm)
Total accumulated
ETr = 359,1 mm, (n=2, ETr = 303,9 mm)
Calculate runoff
Q = P - ETr, = 490 mm – 359,9 mm, = 130,1 mm of runoff, (n = 2, 186,1 mm runoff)
Calculate % Q / P
% = 130,1 mm / 490 mm, = 26,6 % of the precipitation is transformed to runoff, (n = 2, 38 %)
This is more than the 4,3 % of the results of Example 2 for the same temperature range.
28
D R A F T V5b
Conclusions from the Appendices Examples (5 through 9)
The examples were again only simple demonstrations of the concepts and issues involving infiltration and ET.
The importance of time scale and a rough partitioning of wet and dry periods has been demonstrated. Lower
temperatures as demonstrated in Example 2 will reduce ETr and increase runoff approaching the level of the
infiltration models in these notes.
The evaluation of the individual months verifies that as precipitation depths drop and temperatures rise, ETr will
approach the precipitation value. Thus, the reason for having the P/L > 0,316 ratio for the Turc method. If the
ratio < 0,316, ETr is assumed to be equal to the precipitation depth.
This is often assumed for desert conditions when using ETr methods. The assumption is that all the “low depth”
rainfall will infiltrate and replenish the soil moisture. It may be the case for sandy deserts. However, we have
shown with the infiltration models that this is not necessarily the case when plant cover is sparse and when the
soils/rock that are typical for a region are relatively impermeable.
Finally, if “n” = 2 in the equation for “L” in Turc’s equation, the results in terms of the runoff to precipitation
ratio will be higher. These simple examples are not necessarily conclusive, but they do indicate that the Turc
method will better correlate with RCN-based infiltration methods if “n” = 2. In particular, consider the
accumulated Ia (61,5 mm) in the infiltration model of Example 8 of Appendix C. Assume that Ia is not a surface
loss but actually infiltrates into the soil to contribute to the soil moisture and eventually ETr. Then the runoff in
Example 8 will actually be 182,4 mm (244 mm - 61,5 mm) and essentially equal to the runoff in Example 9 of
Appendix D (186,1 mm). In a similar fashion, assume that Ia does not contribute to ETr but to runoff. Add the Ia
to the same ETr (186,1 mm) of Appendix D and we get 247,7 mm.
FIN
29

Tarija Bolivia Climate Part 2 V5b

Uploaded by

Copyright:

Available Formats

Tarija Bolivia Climate Part 2 V5b

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Tarija Bolivia Climate Part 2 V5b

Uploaded by

Copyright:

Available Formats

D R A F T V5b

The purpose and importance of water balance studies

The study region

Figure 1: The four zones of the Department of Tarija. The City of

Figure 3: General climatic data for Tarija

Figure 4: Different siting conditions for a pan evaporator

Figure 5: Two similar cross-sections with the same gage

Figure 6: A 24-hour storm spanning two days. The 24-hour depth is

Figure 7: The accumulated vapor in a plastic bag enclosing a plant is the

Part 3: Precipitation (P), infiltration (I), and evapotranspiration (ET)

Figure 8: Ground water is sometimes a source of soil moisture.

Figure 9: Two rainfall events having different areal

The concept of “effective” precipitation and streamflow

Figure 11: Composition of soils. Relative density (specific

There are three types of water in soils depending on the primary

The forces present in soil-water matrix are important.

Figure 16: In contrast to the general conditions shown in Figure 15

Figure 18: Urban infiltration conditions generally represent an impermeable

Figure 19: Table 2 from Fisher, et al

Figure 20: Remote sensing concepts (CRISP)

Figure 21: Remote sensing capture of sensible and

Figure 22: Table 1 Remote sensing models from Liou

Part 4: Simple Examples

Average Precipitation Years on Record: 15

Average Number of Rainy Days Years on Record: 15

Q = (P – Ia)2 / [(P – Ia) + S] (5a)

Annual runoff based on average annual rainfall

Initial abstraction “Ia”, the new “demon” issue introduced in Example 3

Figure 23: Climatic conditions (circle) for Tarija and

References used for these notes

You might also like