ww.sciencedirect.com

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 0

Available online at w

ScienceDirect

http: / /www.elsevier .com/locate/biombioe

Protocol to compensate net evaporation and netprecipitation in open-pond microalgal massivecultures and permit maximal steady-stateproductivities

Mishka Talent a,b, Greg Burgess a,*, Javier G. Fernandez-Velasco a,c,1

aResearch School of Engineering, The Australian National University (ANU), Canberra, ACT 0200, Australiab Fenner School of Environment and Society, ANU, AustraliacPhotobioenergetics Group, Research School of Biology (RSB), ANU, Australia

a r t i c l e i n f o

Article history:

Received 29 February 2012

Received in revised form

3 March 2014

Accepted 12 March 2014

Available online 8 April 2014

Keywords:

Microalgae

Biomass

Bioenergy

Evaporation

Salinity

Ponds

* Corresponding author. Tel.: þ612 6125 5650E-mail address: greg.burgess@anu.edu.au

1 Present address: Departamento de Cienchttp://dx.doi.org/10.1016/j.biombioe.2014.03.0961-9534/ª 2014 Elsevier Ltd. All rights rese

a b s t r a c t

Commercial cultivation of microalgae in open ponds can be strongly limited by evapora-

tion. This is particularly the case in highly evaporative areas where production plants are

likely to be located. We present a mathematical model to calculate native solute and

exogenous nutrient concentrations as affected by net evaporation or net precipitation. The

model takes into account the periodic compensatory addition of new feedstock water and/

or removal of rain-originated excess water and the eventual recycling of the culture me-

dium. We present a management protocol in which, for a wide variety of climates, it is

possible to stabilise the native solute concentrations and minimise the exogenous nutrient

washout. The protocol includes harvesting a minimum of 10% of the pond volume per day

and replacing this volume and the evaporated water with new feedstock water. We test the

approach against a 9-year daily weather data set for a locality with 1740 mm.y�1 average

net evaporation. We find that the native solute concentrations can be maintained between

1.2 and 1.5 times the feedstock water values, while the daily washout of exogenous nu-

trients is �3% and the pond volume is kept nearly constant. This protocol should prove

particularly useful for the exploitation of marine microalgae growing in open ponds with

brackish or seawater-based growth medium.

ª 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Solar radiation can be harvested as a sustainable source of

energy by exploiting natural photosynthesis. Part of the visible

light energy is stored in biomass, which can be used to

.(G. Burgess).ias Basicas. Universidad019rved.

produce biofuels or a variety of industrial feedstocks. Diverse

“energy crops” are available [1] but, noticeably, some micro-

algae have the highest productivities of biomass and oil [2].

Because of their high productivities, these organisms should

be the best “energy crops” [3] and become the basis for a new

“agriculture”. Furthermore, microalgae display a variety of

Nacional de Lujan. Argentina.

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 082

advantages over classical crops, among others: (a) They do not

need the supporting properties of a soil, and some can grow in

sea or saline water, thus not competing for land or fresh

water; (b) Their culture is hydraulic in nature, which allows

the inoculation and harvest of an extended culture from a

single pumping point and the continuous monitoring and

provision of nutrients; (c) They can display year-round

growth, whereas classical crops are productive only during a

few months per year; (d) Microalgal growth can be fertilized

with CO2 derived from power plant flue gas or from the

extraction of fossil oil and gas, or with CO2 from soluble car-

bonates [3e5]. Other advantages of marine microalgae, when

used in terrestrial industrial setups are: (a) Seawater has

minimal sediment content in comparisonwith rivers. (b) Most

non-marine algae and animals cannot tolerate seawater,

which reduces contamination effects from the terrestrial

surrounding environment. (c) Because most marine micro-

algae cannot tolerate a fresh water environment, any acci-

dental spillage will have a lower environmental impact on

neighbouring ecosystems.

A particular feature of microalgal industrial cultures is that

all the nutrients needed for biomass generation are provided

by the feedstock water and by the addition of fertilizers [6],

here called “exogenous nutrients”.

Microalgal production facilities should be placed in areas of

adequate insolation, moderate temperatures, and easy pro-

vision of water. But also, very importantly for the rational use

of soil resources, they should be located in classically non-

productive land, which typically involves areas of high evap-

oration and low rainfall. Moreover, the extended microalgal

cultivation infrastructure has to be simple and cheap, which

suggests the use of open ponds [7,8]. This however implies

significant water loss through evaporation, in addition to the

water removed for algal harvesting (which can be partly

recycled). The addition of new feedstock water to compensate

evaporation will result in increasing solute concentrations,

which if not managed may ultimately force the replacement

of the entire growth medium [9].

We present a mathematical model to predict the native

solute and exogenous nutrient concentrations as affected by

net evaporation or precipitation, initially utilising an idealised

climate with constant daily conditions, manipulating the pe-

riodic input of feedstock water or removal of growth medium

in combination with the harvesting regime. This model using

constant daily weather is used to demonstrate the role of

those variables, however it can also be directly applied to

short-term periods of steady conditions. For a wide range of

idealised climates from 4000 mm.y�1 net precipitation to

4000 mm.y�1 net evaporation, it is possible to stabilise the

native solute concentrations even with the high level of water

recycling which is required to minimise exogenous nutrient

washout. The case of net precipitation is considered as, even

in arid areas, such conditions will periodically occur andmust

bemanaged. Themodel assumes the harvest of aminimumof

10% of the pond starting volume per day, and the addition of

new feedstock water to compensate both for harvesting and

net evaporation. Growth curves of Nannochloropsis salina

measured experimentally under laboratory conditions are

used in the model. We then test the approach against a 9 year

long daily weather data set for a locality with 1740mm.y�1 net

evaporation, which includes periods of both severe evapora-

tion and precipitation. We find that by daily manipulation of

the plant operating parameters, the native solute concentra-

tions can bemaintained between 1.2 and 1.5 times the starting

values, while the daily washout of exogenous nutrients is�3%

and the pond height increases to at most 20% higher than the

starting value. The system, for the tested weather conditions,

would ideally require the selection or genetic construction of a

microalgal strain capable of proliferating well in a range of

1.0e1.5 relative feedstock solute concentrations.

2. Materials and methods

2.1. Growth of N. salina and outdoors growthsimulation

The marine unicellular microalga N. salina [10], obtained as

strain CS-190, CSIRO Collection of Living Microalgae, Hobart,

Australia, was grown in f2 medium, a half concentration of f

medium [6], prepared with filtered natural seawater (Sydney,

Australia). Cultures were performed in 2 L conical flasks

containing 1.5 L of culture at 25 � 2 �C, magnetically stirred,

continuously illuminated with an average photon irradiance

at the culture surface level of 150 mmol photons.m�2.s�1, and

continuously bubbled with a 5% volume fraction of CO2 in

humidified air, at a flow rate of 0.7 L.h�1. Biomass concen-

tration was evaluated turbidimetrically at 750 nm [11]. In the

model, growth in outdoor conditions was simulated by inter-

spacing 16 h of no growth (i.e. “night conditions”) per each 8 h

of growth (i.e. “day conditions”) using the actual growth curve

obtained under continuous light as described above. This

simulation was subsequently modified to represent the

pseudo steady-state growth in which a culture growing in the

late linear phase is partially harvested each day by the with-

drawal of 10% of the total volume. An 8 h period of linear

growth (maximal rate) is followed by a harvest and addition of

new feedstock water that brings the volume to the standard

value (e.g. equivalent to 300mmpond depth) and causes a 10%

dilution. This dilution is followed by a period of no growth.

The three phases, totalling 24 h, were repeated for many cy-

cles representing the pseudo steady-state culture.

2.2. Model assumptions

2.2.1. Pond type and mixingWe consider a standard pond depth of 300mm.We impose the

general requirement that the mixing mechanism for the open

pond should be effective for assuring the homogenization

through the pond of: a) changes in solute concentration due to

evaporation or precipitation, b) addition of new feedstock

water, and c) addition of recycled water. Raceway ponds with

paddle wheel mixing are the most common commercial

configuration [4], however our analysis is also applicable to

other geometries andmixing devices. In applying this study to

a real system, if mixing significantly increases evaporation

above the local pan evaporation rate, the model is simply

adjusted accordingly. As a variety of evaporation rates have

been evaluated, including extreme ones, the general results of

the model will remain applicable. Further refinement of the

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 0 83

model goes beyond of the purpose of this paper, which is to

introduce amethodology for modelling andmanaging salinity

levels in open ponds. The analysis is not applicable to stag-

nant, unmixed ponds.

2.2.2. Pond evaporation ratesAs noted above, where actual weather data is used, the pond

evaporation rate is set to the measured pan evaporation rate

for the locality.

The rate of evaporation of saline water under given

ambient conditions (temperature, humidity, wind speed) is

affected by the salinity concentration and the surface area of

the body of water [12,13]. These effects are also not considered

in this paper.

2.2.3. Microalgal growth rateWhere growth curves for a microalgal culture subjected to the

regime of harvesting, recycling, and addition of feedstock

water are presented, thesemake no allowance for variation in

growth rates with salinity levels, but only to the removal and

regrowth of the culture. A more detailed model for a specific

microalgal strain should incorporate its growth rate as a

function of the pond salinity level, measured under outdoor

conditions.

This simulation derived from data drawn from laboratory

conditions is not intended to represent actual commercial

production, but only to provide an image of the behaviour of a

culture manipulated as proposed here. The growth kinetics

measured turbidimetrically describes the shape of the ki-

netics, but is not used to estimate absolute biomass

concentrations.

In an open pond, maximal microalgal productivity is

attained with the addition of concentrated CO2. Rather than

being fixed, the rate of CO2 injection would be in accordance

with the photosynthesis rate, as determined by pH

measurement.

2.2.4. Culture harvest rateThe simulation assumes a 10%daily harvest rate (i.e. 10% of the

total volumeof theponds is removed forharvestingof thealgae,

with some fraction of this possibly being recycled back into the

ponds). Some commercial systems use a significantly higher

rate (e.g. 30%) e the effect of this is discussed in Section 4.6.

2.2.5. Feedstock solute concentrationThe feedstock native solute concentration (salinity) is

assumed to remain constant over time (in practice some

feedstock sources may have varying salinity, however this is

beyond the scope of the current model). All native solute

concentrations are given relative to the feedstock

concentration.

3. Theory

3.1. Simulation of solute concentration dependence on,and compensation for, net evaporation and precipitation

The simulations were generated in Microsoft Excel� using

Visual Basic�, version 6.5.

From the universal dilution formula [14], by expressing the

pond volume as a function of its initial height (hp) and the net

evaporation or net precipitation per day as a water height (he:

positive for net evaporation), the actual concentration of any

solute [Si] in the water at time t (in days) contained by a pond

that is subjected to net evaporation or net precipitation can be

simply written as:

Si½ �t ¼hp

hp � the

� � Si½ �0 (1)

where [Si]0 is the concentration of any native solute in the

original feedstock water or any exogenous nutrient added at

time zero.

Equation (1) can be generalised to allow for water removal

or culture harvesting, and the addition of fresh feedstock

water or rain, and partial or full recycling:

Si½ �t ¼hp � he � hw þ hr

� � hphp�he

� �Si½ �t�1 þ hw � hr þ heð Þ Si½ �0

� �

hp(2)

where hw is the height of water withdrawn each day either for

harvesting and/or elimination of excess culture volume due to

net precipitation; hr is the height of harvested culture recycled

to the pond each day; he is the daily evaporation or precipi-

tation and also the height of feedstock water added, or rain

removed each day to exactly compensate that daily evapora-

tion or precipitation, and [Si]t�1 is the concentration of any

solute in the pond the previous day. To generate the solute

concentration time curves this formula is applied iteratively.

The concentration of any solute [Si]t is calculated after har-

vesting and addition or removal of water tomaintain the pond

height (hp) at a constant level. The model uses the simplifi-

cation that all processes (evaporation, harvesting, addition

and removal of water) take place over a short period each day.

For the case of exogenous nutrients [Si]0 in Equation (2) is

set equal to zero, because there are no exogenous nutrients in

the feedstock water. We do not allow for the change in con-

centration of exogenous nutrients due to assimilation by the

microalgae; i.e. in themodel exogenous nutrients are lost only

when the culture is withdrawnwithout full recycling. In a real

system, however, (exogenous) nutrients are assimilated by

growing microalgae. As the rate of assimilation increases, the

amount of nutrients lost when harvesting the culturemedium

decreases. Thus, our calculation of exogenous nutrients

washout, which does not take into account microalgal

assimilation, indicates the absolute maximal loss. Whereas

mostmacronutrients will be faster assimilated by the growing

cells, the micronutrients will be only partially consumed and

the estimation of exogenous nutrient loss could be applied at

least to them.

4. Results

4.1. Change of pond solute concentration in idealisedweather conditions

Sustained net evaporation or precipitation can strongly affect

the microalgal growth medium composition in open ponds

[7,15]. Here we develop simulations to quantify these effects.

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 084

We consider the changes of native feedstock water solute

(“salts”) concentration (“salinity”) and exogenous nutrient

concentration due to net evaporation and net precipitation

and how those changes can be affected or governed by addi-

tions of new feedstock water or removal of growth medium

from the pond. We do not attempt to account for the natural

depletion of native or exogenous nutrients due to their uptake

by the cells during growth; as a compensatory re-addition of

nutrients is always required.

Fig. 1 shows the kinetics of solutes concentration in open

ponds as a function of constant net evaporation or precipita-

tion, in the absence or presence of daily manipulation to

compensate pond volume changes. The application of Equa-

tion (2) shows that the feedstock native solute and exogenous

nutrient concentrations attained by the culture in the absence

of manipulation after (e.g.) 11 days deviate from their original

levels at the beginning of the culture. The pond height de-

creases or increases (not shown) under evaporation or pre-

cipitation conditions, leading to monotonic increases or

decreases in solute concentration (the solute concentration of

rain water is taken as being zero). The concentration of any

exogenous nutrients not assimilated by the algae follows an

identical pattern to the solute concentration.

In order to stabilise the pond volume, a regime of daily

additions of new feedstock water or removal of growth me-

dium from the pond is here employed for periods of net

evaporation or precipitation, respectively. For the case of net

evaporation, curves b, d and f in Fig. 1 show that the

compensatory daily addition of feedstock water retards the

increase in the solute concentration (compare with control

curves a, c and e, respectively). With this protocol, the

Fig. 1 e Concentration of native solutes or exogenous

nutrients versus time. Curves a, c, e, g, i, k and m

correspond to climates of annual net evaporation of 3000,

1000, 500 and 0 mm and net precipitation of 500, 1000 and

3000 mm, respectively, with no manipulation of the pond

volume. Curves b, d, f, and h, and curves j, l and n are the

corresponding curves when daily addition of feedstock

water, or removal of growth medium, is used to keep the

pond volume constant. Curve o represents a simulated

growth curve of N. salina.

concentration of exogenous nutrients not assimilated by the

algae remains constant in the net evaporation regime.

A simulated growth curve for N. salina in an 8 h light/16 h

dark regime is also shown in Fig. 1 (curve o). We adopt this

growth curve as a generic model of microalgal growth.

However, while a daily removal of growthmedium equal to

the daily net precipitation maintains the pond volume con-

stant, the dilution with rain of the original solutes (feedstock

or exogenous) is exacerbated (compare curves j, l, and n with

curves i, k, and m, respectively). This is because the diluting

effect of the rain is higher when acting on the constant orig-

inal pond volume than on a continuously increasing one.

This analysis indicates that, for any givenmicroalgal strain

and growth conditions, the maximum length of time that is

allowed in order for the culture to achieve the highest possible

cell density in the linear growth phase, before a total or first

partial harvest is performed, will be determined by its typical

growth kinetics and “salinity” tolerance, and the rate of net

evaporation or net precipitation in a given location.

4.2. Periodic partial harvest combined with newfeedstock water re-additions stabilises solute concentrationsagainst net evaporation in idealised weather conditions

Once an optimum cell density is reached through growth, a

programme of periodic (e.g. daily) partial harvesting should

follow involving the withdrawal of a fraction of the total vol-

ume (e.g. 10%) and the separation of the cells from the growth

medium, e.g. by centrifugation. In ourmodel this begins at day

11 (Fig. 2). The corresponding supernatant (the growth me-

dium separated from the cells) can be discarded or recycled.

An identical amount of new feedstock water needs to

replace the supernatant discarded in order to maintain the

pond volume and allow the long-term exploitation of the

culture. This brings two consequences. Firstly, the concen-

tration of native solutes that had been changed by net evap-

oration is brought back towards its starting value by the

contribution of the new feedstock water. Secondly, the con-

centration of exogenous nutrients, which have not yet been

assimilated by the algae, is diluted.

Fig. 2A shows those effects for the case of a location with

an average evaporation of 3000 mm.y�1 (corresponding to a

severely arid zone) in an idealised climate of constant daily

conditions. Curves a (control) and b (daily evaporation exactly

compensated with addition of new feedstock water) from

Fig. 1 are presented again in Fig. 2A as references. Curve c in

Fig. 2A shows the evolution of the concentration of the

feedstock-water originated solutes during the complete pro-

cess of biomass growth since inoculation (time zero) and the

pseudo steady-state process resulting from growth and daily

harvesting (10%) which starts after an optimal cell density is

reached. This allows the maximal achievable stabilisation of

native solute concentration whilst maintaining a constant

pond volume. During the period of biomass build-up the

feedstock-water originated solutes concentration increases

linearly (see Section 3.1). However, when the daily harvesting

begins with feedstock water re-addition (no recycling), the

solute concentrations decrease to finally stabilise at a level

1.24 times higher than the concentrations in the original

feedstock water.

C D

A B

Fig. 2 e Native feedstock water solute and exogenous nutrient concentrations during the microalgal culture, with constant

net evaporation and a variety of management conditions. A) Growth medium is not recycled. 3000 mm.yL1 net evaporation

a) Native solute or exogenous nutrients concentration with no water additions or harvesting. b) Feedstock native solute

concentration with daily additions of new feedstock water to compensate the average net evaporation; no harvesting. c)

Feedstock solute concentration where two regimes are applied: for the first 11 days, new feedstock water is daily added as

for curve b; subsequently, during the pseudo steady-state phase generation of biomass combined with daily partial harvest

(10%), feedstock water is added to compensate for evaporation and also to replace the volume that is withdrawn daily for

harvesting. The supernatant of this harvested volume is discarded. d) Exogenous nutrient concentration resultant from the

same protocol described for curve c; nutrients are added at time zero. e) Simulated growth curve (biomass concentration) of

N. salina showing the initial biomass build-up (first 11 days with no harvesting), after which growth is combined with daily

partial harvesting. B) Growth medium is recycled in various percentages. 3000 mm.yL1 net evaporation. Curves aee) The

same as Panel A, curve c, but the withdrawn volume for harvest is recycled in the following percentages: a) 100%, b) 95%, c)

80%, d) 50% and e) 0%. fej) Exogenous nutrient concentration resulting from various harvesting regimes in which the

withdrawn volume for harvest is recycled in the following percentages: f) 100%, g) 95%, h) 80%, i) 50% and j) 0%. Curves e, j

and k in Panel B are identical to curves c, d and e in Panel A. C) Growth medium is not recycled. Curves aee) The same as

Panel A, curve c, but corresponding to 500, 1000, 2000, 3000 and 4000 mm.yL1 net evaporation, respectively (curve d is

identical to curve c in Panel A). fej) Exogenous nutrient concentrations corresponding to regimes a-e, respectively; curves

are only slightly different and are unresolved. k) Simulated growth curve as e in panel A. D) Growth medium is 80% recycled

in all cases. aee) The same as Panel B, curve c, but corresponding to the same conditions as curves a-e in Panel C. fej)

Exogenous nutrient concentrations corresponding to regimes a-e. k) Simulated growth curve as e in panel A.

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 0 85

The kinetics of concentration of exogenous nutrients has a

different pattern, as determined by Equation (2). During the

biomass build-up in a regime of net evaporation their con-

centration remains constant (Fig. 2A, curve d), because the

total pond volume is kept constant through the daily additions

of feedstockwater. However, awashout of the initially present

exogenous nutrients starts during the periodic harvesting in

the pseudo steady-state phase (the nutrients assimilated by

themicroalgae are not considered; in practice nutrients would

need to be added to compensate both for microalgal uptake

and the washout due to the addition of feedstock water).

In order to save exogenous nutrients and minimize the

energy required for pumping of feedstock water the harvest-

derived supernatant could be partially recycled. Fig. 2B

shows a simulation for the same climate of 3000 mm.y�1 of

net evaporation with varying levels of recycling. Whereas

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 086

water recycling from the harvested volume produces a higher

steady-state concentration of feedstock water solutes, it

minimizes exogenous nutrient washout. For example, after 50

days of total culture in a regime of 80% water recycling the

exogenous nutrient washout is �56%, mostly independent of

net evaporation as indicated above; whereas the solute con-

centration is 1.85 times the starting value (i.e. the concentra-

tion in the feedstock water).

Local climate conditions and recycling rate have a strong

impact on the kinetics of the native solute and exogenous

nutrient concentrations and are shown in Fig. 2C and D

respectively. Identical simulations considering awide range of

net evaporations show that the feedstock water solute con-

centration stabilises closer to the original level the lower the

net evaporation. For an extreme case of net evaporation of

4000 mm.y�1, Fig. 2C shows that the solute concentrations

reach a peak of 1.37 at day 11 to finally stabilise at a level 1.32

times higher than in the feedstock water. The kinetics of the

exogenous nutrient washout is essentially independent of the

net evaporation regime as it results mainly from the per-

centage of daily culture withdrawal. For 10% recycling the

daily washout is w10% (Fig. 3), whereas for 80% recycling it is

w2% (Figs. 2D and 3), in each case nearly independent of the

net evaporation. While this benefit is achieved at the expense

of higher pseudo steady-state native solute concentrations,

these do not exceed 2.5 times the original concentration even

with the extreme climate of 4000 mm.y�1.

Fig. 3 e Summary of the effects of different climates of

idealised constant daily weather and water recycling

regimes on the native solute (salt) concentration and the

daily rate of exogenous nutrient loss. The native solute

concentration values are computed after 90 days from

inoculation. The rate of exogenous nutrient loss due only

to the exchange of water between the pond and the

feedstock water reservoir is expressed as the daily

percentage loss for a pond 300 mm deep and a daily

harvesting regime of 10%. Solid curves a through k

correspond to 4000, 3000, 2000, 1000, 500 and 0 mm of

annual net evaporation, and 500, 1000, 2000, 3000 and

4000 mm of annual net precipitation, respectively. The

dashed curve symbolizes the theoretical case for a climate

of 4000 mm annual net evaporation after 1 year from

inoculation and continuous pond operation. The recycling

regimes (0e100%) are depicted by the straight lines. The

dotted line indicates the limitation of recycling values

under net precipitation. The shaded area indicates

climates and recycling conditions in which transient

values of native solute concentration with values above the

computed ones in the figure take place at the beginning of

the production phase.

4.3. Periodic partial harvest combined with newfeedstock water re-additions stabilises solute concentrationsagainst net precipitation

Net precipitation can also affect industrial cultures of micro-

algae in open ponds, and must be considered even for regions

with high annual evaporation. Rain causes both an increase of

the pond volume and a dilution of the feedstock water native

solutes and the exogenous nutrients (Fig. 1). To maintain the

pond volume constant and the feedstock native solutes sta-

bilized, the growth medium could be removed daily from the

pond in an amount equal to the net precipitation and new

feedstock water added, with an equivalent protocol as the one

used for net evaporation conditions in Fig. 2, also applying

Equation (2). To achieve this, part of the harvested volume is

discarded in an amount equal to the rain volume, with the

consequence that recycling must be lower than 100%. There-

fore, a substantial reduction of the dilution of the feedstock

water native solutes in the growth medium occurs, compare

data in Figs. 1 and 3.

During rainy periods, the rate of dilution of exogenous

nutrients increases (not shown) because of the double effect

of disposing growth medium in an amount equal to the rain

plus the harvested volume while maintaining the pond vol-

ume constant by the addition of new feedstock water. The

recycling strategy to save exogenous nutrients already pre-

sented for the case of net evaporation can also be employed

for net precipitation (Fig. 3). However, this saving of exoge-

nous nutrients is achieved at the expense of lower pseudo

steady-state native solutes concentrations, which is the

converse of the case of net evaporation. Furthermore, the

deviation from the original level is much more pronounced

than for the case of a net evaporation of equal absolute value.

For the case of net precipitation, feedstock water solute

concentration stabilises closer to the original level the lower

the net precipitation. For climates between 500 mm.y�1 and

4000 mm.y�1 net precipitation, concentration stabilisation

occurs between 0.96 and 0.66 times the original feedstock

water concentration, respectively (Fig. 3).

The kinetics of the exogenous nutrient washout is essen-

tially independent of the net precipitation regime, as it results

mainly from the percentage of daily culture withdrawal.

Indeed, the daily washout is 9.95% and 9.65% of the prevailing

concentrations for climates of net precipitation of 500mm.y�1

and 4000 mm.y�1, respectively (Fig. 3). While recycling can

reduce this exogenous nutrient washout, the maximum

recycling is limited by the need to remove the precipitated

volume to maintain the pond volume constant. For example,

with an average net precipitation of 3000mm.y�1 themaximal

recycling possible is 73% (Fig. 3).

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 0 87

4.4. Summary of achievable pseudo steady-state nativesolute concentrations and exogenous nutrient loss rates fora variety of recycling regimes

The achievable pseudo steady-state native solute concentra-

tion and the associated rate of exogenous nutrient loss for a

variety of idealised climates with constant conditions are

shown in Fig. 3. A series of climates ranging from net evapo-

ration of 4000 mm y�1 to net precipitation of 4000 mm y�1 are

considered under different water recycling regimes. The

native solute concentration is evaluated after 90 days from

inoculation, as this could be a possible duration of a culture

before it has to be restarted (cf. curve e in Fig. 2A). For strong

net evaporation (e.g. �2000 mm.y�1) and high recycling (e.g.

>50%) the native solute concentration is not stabilised after 90

days, as exemplified in Fig. 2B and D, whereas with low

recycling (e.g.�50%) the native solute concentration stabilises

before 90 days for all climates, as exemplified in Fig. 2B. This

difference is also shown by the dashed curve in Fig. 3 which

represents the same conditions as for curve a in Fig. 3 (i.e.

4000 mm.y�1 net evaporation) but after 1 year of a theoretical

continuous operation of a same culture: both curves concur at

low recycling and profoundly diverge at high recycling.

A

C

B

Fig. 4 e Test of the proposed protocol in an actual climate scenar

concentration with inputs of new feedstock water for a locality w

against 9 years of actual climate data. A) Daily net evaporation

days with net precipitation). B) Pond water depth. Curves A and

of the harvested culture. D) Concentration of native feedstock w

Fig. 3 again shows that increasing net evaporation or pre-

cipitation results in a stronger deviation of the solute con-

centration from its starting point of 1. This can be minimized

by recycling less, but at the expense of a higher rate of exog-

enous nutrient loss; which in the case of 0% recycling reaches

a value close to a daily loss of 10% of the nutrients present at

any time. In the case of net precipitation there is a maximal

level of recycling that can be achieved (Section 4.3). This is

indicated by the dotted line in Fig. 3, which interrupts the

curves of net precipitation.

Fig. 3 also shows that the same percentage of water recy-

cling affects somewhat differently the percentage of exoge-

nous nutrient daily loss for different climates. As derived from

Equation (2), this difference is smaller the higher the recycling.

In net evaporation conditions and with recycling <22% the

concentration of native solutes at the end of the growth period

(before harvesting starts) is higher than the level achieved

after 90 days. In the most extreme case considered here, i.e.

4000 mm.y�1 with 0% recycling, this transient peak (“over-

shoot”) reaches a concentration 3.8% higher than the level

after 90 days (see curve e in Fig. 2C). The shaded area in Fig. 3

indicates those climates and recycling regimes where such

overshoots occur.

D

io. The protocol of modulating native feedstock water solute

ith an average annual net evaporation of 1740mm is tested

(i.e. evaporation e precipitation. Negative values represent

B are explained in the text. C) Percentage of water recycling

ater solutes.

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 088

4.5. Simulations with actual weather data

We tested if the conclusions obtained with simulations using

idealised climates, as described above, also apply under actual

daily weather conditions. The case of Tullakool in New South

Wales, Australia (35�220 S; 144�120 E) was considered, using a 9-

year record of precipitation and pan evaporation from the

Australian Bureau of Meteorology. The net evaporation is

shown in Fig. 4A where regular annual cycles of evaporation

are accompanied by both events of heavy rain (negative de-

flections) and above average evaporation.

We consider a possible management protocol to test

whether we can accommodate these natural variations from

the yearlymean of net evaporation; managing pond feedstock

water solutes concentration by the combined action of

importing new feedstock water and modulating the percent-

age ofwater recycling. For this test protocol a constant harvest

of 30 mm per day is taken, irrespective of the pond water

depth. The target for the water depth is 300 mm, whilst the

target range for the solute concentration is 1.0e1.5 times that

of the feedstock water.

The simulated management of the ponds involves: a)

considering the actual daily net evaporation or precipitation

and b) the highest possible recycling of the harvested volume

water in order to minimize the exogenous nutrient washout.

However, in the simulation the daily percentage of recycling

is kept in the range of 20e90% and is chosen so as to keep the

native solute concentration below 1.5. The limiting value of

salinity of 1.5 is chosen so as to maintain high productivity

[16]. It is assumed that microalgal strains can be collected

from the environment [2] or developed, which are tolerant to

this degree of salinity variation. The rationale for the limits

for the recycling percentages involves aiming to minimize

exogenous nutrient washout but to have at least some

renewal of feedstock water so as to dilute microalgal

secretions.

Fig. 4B shows the consequences of the management pro-

tocol on the pond depth (curve B). A very stable pond depth

with extremes between 300 and 362 mm results. Moreover,

the deviations from the desired value are short-lived, at most

lasting 7 days. For comparison, a more rigid protocol is also

shown (curve A) in which daily feedstock additions are con-

stant, corresponding to the yearly average evaporation of

4.77 mm.d�1 and daily recycling of 70% of the harvested me-

dium. For this inflexiblemanagement protocol the pond depth

(curve A) and the native solute concentrations (not shown)

oscillate widely. Such inflexible management protocols are

therefore not further discussed.

Fig. 4C compares the percentage of water recycling of the

harvested volume. The amount of recycling is daily varied

depending on the actual net evaporation or precipitation.

Relatively long and uninterrupted maximal recycling (90%)

periods coincide with periods of high precipitation or rela-

tively low net evaporation, as in these circumstances the

native solute concentration is maintained well below the

chosen limit of 1.5. During periods of high evaporation the

average recycling diminishes. The high variability in the level

of recycling results from the need to react to daily net pre-

cipitation or evaporation.

When days of significant rain occur in this case of average

annual net evaporation, the recycling strategy to maintain

native solutes close to the feedstock concentrations is oppo-

site to the case of a climate of average net precipitation. In the

example of Fig. 4, during rainy periods, in order to maintain

the native solutes concentration close to 1.5, low inputs of

feedstock water (concentration ¼ 1) and maximal recycling

are required.

Fig. 4D shows the resulting solute concentration levels for

the pond. The salinity initially rises quickly, but then is

maintained between 1.2 and 1.5 times the feedstock solute

concentration for the rest of the entire 9-year period. The

maximum concentration of 1.5 is higher than that which

occurred for constant 3000mm.y�1 evaporation (1.24) as in the

real climate there are periods of evaporation significantly

higher than the average.

The daily exogenous nutrient loss varies between 0.8 and

8.0% of the total content of each exogenous nutrient in a pond

of constant 300 mm height (not shown). The maximal and

minimal loss of exogenous nutrients coincides with times of

minimal and maximal recycling, respectively. Despite this

seasonal variation, the average daily exogenous nutrient loss

is 3.0%, identical to the yearly idealised climate that is depic-

ted in Fig. 3, simply resulting from the premises chosen to

begin the simulation.

4.6. Effect of culture harvest rate

For a production system with a different harvest rate to 10%

the solute and nutrient concentration curves should be

recalculated. However some insight into the effect of different

harvest rates can be obtained by considering the inter-

relationship between harvesting and recycling. The net ef-

fect of harvesting a fraction fH of the pond volume and then

recycling a fraction fR of the harvest is that the pond height is

decreased by an amount Dh given by Equation (3):

Dh ¼ fH�1� fR

�hi (3)

where hi is the height of the pond at the start of the harvesting

and recycling process. It is seen that the harvesting and

recycling rates only affect the results via the amount of water,

which is actually discharged (i.e. the blowdown) from the

system. For example, a regime with 10% harvest rate and 50%

recycling is directly equivalent to one with 30% harvest and

83% recycling.

In general terms, higher harvest rates (for a fixed recycling

percentage) lead to native solute concentrations stabilizing

closer to the feedstock concentration. However, a higher

washout of exogenous nutrients would also occur.

5. Discussion

The protocols described in this papermust be accompanied by

proper industrial management of the microalgal culture to

obtain maximal steady-state productivities. We identify the

following key aspects which are exemplified in curve e in

Fig. 2A.

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 0 89

� Cultures must be exploited at their maximal productivity.

� Cultures must be protected against the problems of pho-

toinhibition (excessive effective irradiation of the cells

[17]); and biological contamination.

Both of these requirements are related to the culture cell

concentration. The maximal areal productivity of the culture

is maintained up to the end of the linear growth phase,

which develops after the cell division exponential phase [18].

Harvesting should occur at the end of the linear phase,

before the decline towards the culture stationary phase, and

it should be only a partial harvest. This allows the mainte-

nance of a permanent and relatively high cell concentration

in the culture, which also minimizes photoinhibition and

biological contamination. A steady-state of high cell con-

centration also avoids the risks associated with a full harvest

and a restart of the culture with lag and exponential

phases [18].

5.1. Model involving idealised climates with identicaldays

The continuous operation of a production pond in a given

climatewill in general be associatedwith the establishment of

higher or lower than feedstock solute concentrations. The

proposed protocol involves exchanges of water between the

pond and the water reservoir, which determine the loss of

exogenous nutrients. The information shown in Fig. 3 can be

used to determine, for a variety of climates of idealized con-

stant daily weather, the quasi steady-state native solutes

concentration after 90 days of inoculation and the resulting

rate of exogenous nutrient loss. For climates with very regular

weather patterns, particularly in short term spans, the data of

Fig. 3 can be directly used.

5.2. Model involving an actual climate

The results in Fig. 4 demonstrate that the application of the

concept of compensating net evaporation or precipitation

with the addition of feedstock water and/or the removal of

growth medium is feasible in a real climate scenario.

The loss of exogenous nutrients is determined solely by net

evaporation or precipitation and the rate of recycling. The

latter is determined by the chosen upper bound of the solute

concentration that the pond may attain, which is most likely

dictated by the microalgae chosen for cultivation.

Interestingly, if the selected microalga tolerates much

higher than normal feedstock water solutes, evaporation of

the growth medium before inoculation could be an excellent

means of minimizing biological contamination. Furthermore,

the pond should be operated as close as practically possible to

the maximum solute concentration that the microalgae can

tolerate in order to reduce exogenous nutrient loss.

5.3. Determination of system operating parameters

The results of the model, as given by Fig. 3, can be used to set

or determine the operating conditions for a given pond sys-

tem. Two of the variables are chosen as fixed and the third is

read as the dependent parameter:

1) A given rate of net evaporation and algal solute concen-

tration tolerance determine the average maximum level of

recycling.

2) A given rate of net evaporation and chosen acceptable

nutrient loss determine the pseudo steady-state solute

concentration.

3) A given algae solute concentration tolerance and chosen

acceptable nutrient loss determine the range of usable

climates.

The selected operating point (or range) will also determine

the rate of loss of exogenous nutrients.

The analysis can be carried out for the selected represen-

tative periods (e.g. summer) of the climate at a real site, by

extrapolating the daily evaporation or precipitation to a yearly

value, after factoring in the increase in evaporation caused by

pond mixing.

5.4. Effect of solute concentration on microalgal growthrate

In the simulation with actual weather data, the solute

concentration increases up to 1.5 the feedstock value. A

number of microalgae, including N. salina, have been found

experimentally to show nearly unchanged growth rates at

salinity levels between 22 and 34 PSU (Practical Salinity

Units) [5] [19], i.e. a range of just over 1.5. However at the

higher salinity levels, which would occur if seawater is the

feedstock (so that the starting point was 35 PSU), growth

rates will tend to decline [19]. As noted previously, a more

detailed model for a specific microalgal strain should

incorporate its growth rate as a function of the pond salinity

level.

6. Conclusion

We have shown that the solute concentration (“salinity”)

analysis and management protocol here presented demon-

strate the possibility of very large scale industrial produc-

tion of biofuels or other materials through microalgal

photosynthesis in areas with strong atmospheric water

unbalances, in particular allowing the exploitation of arid

areas and the use of seawater as the main growth medium

component. The protocol can, with appropriate changes to

the evaporation rate, be applied to any form of open pond

with forced mixing.

Acknowledgements

JGFV dedicates this work to his former teacher at Colegio

Nacional de Buenos Aires, Universidad de Buenos Aires, Prof.

Oscar Perazzo. We thank Dr. John Benemann for useful data

and discussions. This work was supported in part by an ANU-

Linkage Pilot Fund and an RSB-ANU Biotechnology Research

Centre Grant to JGFV and by an RSB-ANUHonours scholarship

to MT.

b i om a s s a n d b i o e n e r g y 6 4 ( 2 0 1 4 ) 8 1e9 090

r e f e r e n c e s

[1] Duke J. Handbook of energy crops [Internet]. PurdueUniversity: Center for New Crops and Plants Products; 1983[cited 2012 Feb 27]. Available from:, http://www.hort.purdue.edu/newcrop/duke_energy/dukeindex.html.

[2] Sheehan J, Dunahay T, Benemann J, Roessler P. A look backat the U.S. Department of Energy’s Aquatic Species Program:biodiesel from algae. Golden, CO: National Renewable EnergyLaboratory; 1998. p. 294. NREL/TP-580-24190.

[3] Benemann J, Pedroni P. Biofixation of fossil CO2 bymicroalgae for greenhouse gas abatement. In: Amadei A,editor. Encylopedia of hydrocarbons. Rome, Italy: Institutodella Enciclopedia Italiana; 2005.

[4] Benemann J. Microalgae for biofuels and animal feeds.Energies 2013;6(11):5869e86.

[5] Brennan L, Owende P. Biofuels from microalgae e a review oftechnologies for production, processing, and extractions ofbiofuels and co-products. Renew Sustain Energ Rev2010;14(2):557e77.

[6] Andersen R. Algal culturing techniques. Burlington, MA:Elsevier Academic Press; 2005.

[7] Richmond A, editor. Handbook of microalgal culture e

biotechnology and applied phycology. Oxford: BlackwellPublishing; 2004.

[8] Borowitzka M. Culturing microalgae in outdoor ponds. In:Andersen R, editor. Algal culturing techniques. Burlington,MA: Elsevier Academic Press; 2005.

[9] Borowitzka MA, Moheimani NR. Sustainable biofuels fromalgae. Mitig Adapt Strategies Global Change 2013;18(1):13e35.

[10] Graham L, Wilcox L. Algae. Upper Saddle River, NJ: PrenticeHall; 2000.

[11] Tassan S, Ferrari G. An alternative approach to absorptionmeasurements of aquatic particles retained on filters. LimnolOceanogr 1995;40(8):1358e68.

[12] Leaney F, Christen EW. Evaluating basin leakage rate,disposal capacity and plume development; 2000. CSIRO Landand Water 17/10.

[13] Ali H, Madramootoo CA, Abdel Gwad S. Evaporation model ofLake Qaroun as influenced by lake salinity. Irrig Drain2001;50(1):9e17.

[14] Harris D. Quantitative chemical analysis. New York:Freeman; 1991.

[15] Wheaton F. Aquacultural engineering. New York: John Wiley& Sons; 1977.

[16] McLachlan J. The effect of salinity on growth and chlorophyllcontent in representative classes of unicellular marine algae.Can J Microbiol 1961;7(3):399e406.

[17] Long SP, Humphries S, Falkowski PG. Photoinhibition ofphotosynthesis in nature. Annu. Rev Plant Biol1994;45(1):633e62.

[18] Lewin R, editor. Physiology and biochemistry of algae. NewYork: Academic Press; 1962.

[19] Bartley M, BoeingW, Corcoran A, Holguin F, Schaub T. Effectsof salinity on growth and lipid accumulation of biofuelmicroalga Nannochloropsis salina and invading organisms.Biomass Bioenerg 2013;54:83e8.