Approach using the Internet of Things in the Sahel for Smart Irrigation

Fatoumata Thiam

and Cheikhou Thiam

Universit

e Gaston Berger Saint-Louis, Senegal

Universit

e de Thi

es, Thi

es, Senegal

Keywords:

Internet of Things, Irrigation, Artiﬁcial Intelligence.

Abstract:

Nowadays the Internet of Things (IoT) is used in many sectors such as health, military, industry, agriculture

and so on. This technology is considered as a special type of ad hoc network and one of these promising areas

is precision farming where it can provide important support that will facilitate agricultural practices and that

can modernize and replace some of the traditional techniques. In Sahelian area, water is a critical resource due

to short rainy seasons and waste of water in gross irrigation. The main activity in the Niayes area is market

gardening. Urgent measures must be taken to save water reserves during the long dry season. Smart irrigation

can be a solution for water wastage problem. Many techniques have been developed to solve these problems.

In this paper, we can ﬁnd techniques based on evapotranspiration (ET), Soil Water Assessment Tool (SWAT)

and Internet of Things (IoT). This paper aims to propose a solution of smart irrigation based on IoT.

1 INTRODUCTION

In recent years, with the evolution of technology, the

emergence of a new internet of things information

network called the Internet of Things , which is con-

sidered as a special type of ad hoc network whose

nodes are sensors able to harvest and transmit envi-

ronmental data in an autonomous way. The nodes are

randomly dispersed across a geographic area, called

the catchment ﬁeld, which deﬁnes the area of interest

for the captured phenomenon. The captured data is

routed to a node considered a data collection place,

called sink node. The latter can be connected to the

network user via the Internet or a satellite. Thus, the

user can address requests to the other nodes of the

network, specifying the type of data required and har-

vest the environmental data captured through the well

node. Internet applications of objects are numerous.

They include different ﬁelds: agricultural, health, ...

etc. The intelligent irrigation system is an IoT-based

device that is able to automatically deactivate the mo-

tor pump once the soil moisture sensor has reached

the required threshold value. There is a growth in

agricultural products via the Internet of Things. This

solution is a response to the problems of environmen-

tal change in the Sahel and has a signiﬁcant impact on

the agricultural economy.

Population grows at an exponential rate and feed-

ing people is progressively being problematic. Ex-

tra food needs to be produced. People are building

more and agricultural land are being occupied also,

the use of water has signiﬁcantly increased. Valuable

land and water resources required for food production

have become critical.

An urgent need of regulating water utilization is

required. In fact, decreasing water, drying of rivers,

unpredictable environment is very common. In Sahe-

lian area, water is a critical resource due to short rainy

seasons and waste of water in gross irrigation. The

main activity in the Niayes area is market gardening.

To solve that, in agriculture domain, sensors are de-

ployed at suitable locations to monitor crop growth

and water use.

In this article we propose an approach that consists

of setting up a network of sensors for intelligent irri-

gation and thus an efﬁcient use of water. This study

will be discussing on how the precision irrigation is

implementing via many technicals. We will next see

those technicals and present some example of works.

This article is organized like following: we will intro-

duce ﬁrst this study, secondly. The section 2 presents

some background concepts. Section 3 presents sev-

eral studies in the domain. In the section 4 we present

our approach for optimizing irrigation. The section 5

presents results which are the comparison of different

approaches. Finally, the last section summarizes the

contributions.

Thiam, F. and Thiam, C.

Approach using the Internet of Things in the Sahel for Smart Irrigation.

DOI: 10.5220/0009800601690174

In Proceedings of the 9th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS 2020), pages 169-174

ISBN: 978-989-758-418-3

169

2 BACKGROUND CONCEPTS

2.1 Irrigation Paradigm in the Niayes

Area

The world’s population is expected to exceed nine bil-

lion by 2050, three times more than in 1950 (Angus

and Butler, 2014)(Bricas and Seck, 2004). The prob-

lem of global food security is growing. The rural exo-

dus leads to an agglomeration form in the urban areas,

which have become in a few years the place of life, of

more than half of the population. The planet is facing

a problem of meeting the food needs of its population.

Food security in urban areas is even more worry-

ing in the cities of the South, already facing a high

rate of poverty. In this context, urban and peri-urban

agriculture (UPA) is developing, occupying an essen-

tial place in the urban food supply.

In Senegal, UPA activities are concentrated in the

Niayes area Figure 1. The Niayes is a series of small

valleys between dunes along the northern coastline

of Senegal. Since its appearance in Senegal in 1937,

market gardening has generally evolved both in terms

of surfaces used and yields. Nevertheless, the demand

is not satisﬁed, and Senegal continues to import veg-

etables.

Figure 1: Niaye area in Senegal (translation).

Indeed with the scarcity of rain, the main source

of water supply that constituted the semi-superﬁcial

aquifers Niayes has shown its limits. As a result,

wastewater has become an attractive alternative for

market gardeners. However, at the bacteriological

level, the coliform and Powerful streptococci concen-

trations are well above the WHO (World Health Or-

ganization) guideline (Niang, 1996).

There are three main constraints identiﬁed by

farmers (Ba and Cantoreggi, 2018):

• land insecurity (74%);

• lack of water for irrigation (62%);

• and salinization of water and soil (27%).

In view of all these problems, a water saving solu-

tion is needed

2.2 Mathematical Models

It is important to optimize water utilization in agricul-

ture to avoid water stress in the future. The scheduling

and management of irrigation water are used to pre-

vent over-irrigation and water wastage. Many tech-

niques are based on studying the plant water require-

ment and other on mathematical model, algorithms,

etc.

Optimized-irrigation can be deﬁned as applying

the right amount of water to the plant, to allow it to

grow without over-irrigating or under-irrigating. In-

deed, over-irrigation can lead to leaching of soil nu-

trients, while under-irrigation prevents the plant from

growing properly.

Smart-irrigation is a set of hardware and software

to automate and optimize irrigation for a good evolu-

tion of the crop. For an optimized irrigation schedule,

there are two components to consider (George, 2017):

1. Determine the crop water requirements (CWR);

2. Estimate the right time to water the plants.

To determine the CWR, there is a parameter called ET

standing for evapotranspiration. The ET is the sum of

crop transpiration and evaporation of soil water con-

tent. The CWR is the same as ET. To determine ET,

the FAO (Food and Agriculture Organization of the

United Nations) have stood two methods (Savva and

Frenken, 2002) :

• Fao Penman-Monteith method is an equation that

require climate parameters to determine the ET

;

• Pan evapotranspiration is a measurement that uses

a pan evaporimeter and parameters like wind

speed, humidity, temperature and the sunshine to

calculate the ET

is the evapotranspiration used as reference to cal-

culate other crops ET wich is ET

equation 1. The

reference crop used for ET

is grass.

Fao Penman-Monteith equation.

0.408δ(R

− G) + γ

900

T +273

− e

)

δ + γ(1 + 0.34u

)

(1)

= Reference evapotranspiration (mm/day)

= Net radistion atr the crop surface (MJ/m

)

SMARTGREENS 2020 - 9th International Conference on Smart Cities and Green ICT Systems

170

G = Soil ﬂux density (MJ/m

)

T = Mean daily air temperature at 2m height (° C)

= Wind speed at 2m hieghr (m/sec)

= Saturation vapour pressure (kPa)

= Actual vapour pressure (kPa)

The FAO Penman-monteith (Savva and Frenken,

2002) is used to determine the ET

and then the IR,

Irrigation requirement equation 2.

IR = ET

− (Pe − Ge +Wb) + LR (2)

IR= Irrigation Requirement (mm)

= Crop Evapotranspiration (mm)

Pe= effective dependable rainfall (mm)

Ge= Groundwater contribution from water table

(mm)

W b = Water stored in the soil at the beginning of each

period (mm)

LR = Leaching Requirement (mm)

It is important to distinct CWR and IR. In fact, for

an effective irrigation, more parameters like the cli-

mate, the crop type and the growth stage of the crops,

need to be taken into consideration to attain plant wa-

ter need.

2.3 Computer Simulation Model for

Natural Water Resources

Computer science helps to model efﬁcient water use

in crop ﬁeld. Powerful evaluation and simulation

tools are made to help water usage planning. To re-

ﬂect actual values of ﬁeld parameters, these models

have to ﬁt in real time system to avoid ineffective

performance of irrigation schedule. In addition, crop

growth, depend closely to irrigation frequency and

amount of water supplied to the plants. So, it is cru-

cial to have a model to produce a good watering plan.

There are many tools used to model water use in crop

ﬁeld, for example (Chen et al., 2018):

• The Soil Water Assessment Tool (SWAT) for wa-

tershed scale model, used to simulate the quality

and quantity of surface and ground water and pre-

dict the environmental impact of land use, land

management practices, and climate change.

• MODFLOW wich is a program used by hydro-

geologists to simulate the ﬂow of groundwater

through aquifers.

• Decision Support System for Agrotechnology

Transfer (DSSAT) is software application pro-

gram that comprises dynamic crop growth simu-

lation models for over 40 crops.

SWAT model is widely used for managing water re-

sources.

2.4 Internet of Things - IoT

ICTs has a signiﬁcant impact on the concept of smart

agriculture. Smart-agriculture is the use of computer

devices, in the farming process to optimize usage of

natural resources, limit human labor and have a good

proﬁtability.

The Internet of Things (IoT), in particular Wire-

less Sensor Networks (WSN) is commonly used in

Smart-agriculture. WSN is a collection of physical

modules called sensors to capture, process and trans-

mit information in the environment in which they are

deployed. Sensors are deployed at different soil depth

to sense water ﬂow and help to determine proper time

for water supply. Based on IoT, SMART-Irrigation

can help for water wastage problems, over-irrigation

that can lead to the leaching of nutrients and slowing

down crop growth (George, 2017). Precise irrigation

is a concept of saving water supply to use exactly the

exact amount of water for the plant to develop cor-

rectly. Many parameters are taken in account to deﬁne

a precise irrigation. Some of them are meteorological

data like temperature, humidity, the wind, etc. Some

of those can be obtained by using sensing elements

like soil moisture sensors.

3 RELATED WORK

SMART-Irrigation is the use of technology in the

farming process to automatize certain procedures and

increase yields. Sensor devices are linked by wireless

communication technology to monitor data from the

agricultural ﬁeld. Many studies are focused on using

soil moisture sensors to determinate water supply or

to determine when to irrigate. In the following we

will see some work on Smart-irrigation.

CROPWAT(Savva and Frenken, 2002) is a soft-

ware that can calculate CWR and IR from climatic

and crop data in a particular area. Various tables are

given to estimate the exact value of each parameter of

the Penman-Monteith equation. Meteorological sta-

tions are distributed all through a given country mak-

ing it possible to prepare reference crop evapotran-

spiration maps for a country. Those meteorological

stations are contained in CLIMWAT wich is a soft-

ware that help to obtain value of each parameter of

the Penman-Monteith equation. I

In (Chen et al., 2018), authors created a SWAT

based algorithm. The algorithm simulates manage-

ment allowed depletion (MAD) irrigation schedul-

Approach using the Internet of Things in the Sahel for Smart Irrigation

171

ing by taking in an allowable depletion percentage of

plant available water (PAW). The algorithm also sus-

pends irrigation events after harvest. Weather data are

acquired from the nearby meteorological observation

post. They are then compiled into daily values and

formatted for model input.

In (Myers et al., 2017), the developed system uses

real-time sensor data, weather forecasts, geological

and environmental information to infer the precise

amount of water needed to minimize wastage without

compromising the health and wellbeing of the lawn or

garden. The water supply in the yard is automatically

managed by sensor-actuator nodes, based on seman-

tic inference. The combination of data from multiple

sources with a sensor-actuator system helps to pro-

vide precise irrigation.

In (Mohanraj et al., 2017), the proposed system

automates the irrigation and fertigation using WSN to

detect rainfall intensity. The system is integrated with

irrigation module which uses Penman-Monteith FAO-

56 equation for calculating crop water need. There

are four functionalities carried out by end nodes: es-

timation of CWR, calculation of irrigation period, de-

tection of water discontinuity and monitoring the re-

maining energy of the battery.

Studies have identiﬁed insufﬁciencies in the auto-

irrigation algorithms in the Soil and Water Assess-

ment Tool (SWAT). It is noted a continuation of ir-

rigation during the non-growing period and incapa-

bility to simulate growth stage speciﬁc irrigation. The

CROPWAT program produces a watering calendar for

all the growing season of the crops once. That type of

conﬁguration may not take in account some climatic

change like rain, slopes or unexpected dryness.

In the following part we are aiming to present our

solution of smart-irrigation wich will consider evap-

otranspiration and real time monitoring plant water

need for a precise irrigation.

4 APPROACH

In our proposal, a set of soil moisture sensors are de-

ployed over the land to monitor soil water content.

Then based on deﬁned thresholds, IR (Irrigation Re-

quirement) will be calculated to determine water sup-

ply, in addition of a set of external data sources. The

system should be able to adapt watering plan through

artiﬁcial intelligence function. Parameters like effec-

tive dependable rainfall (Pe) and groundwater contri-

bution (Ge) (Figure 2) will be acquired from external

data sources (OpenWeatherMap.org, ) (wea, ).

Our system will consist of soil moisture sensors,

actuators and sprinklers nodes. The sensors measure-

Figure 2: Flow diagram.

Table 1: Crop evapotranspiration for tomato crop in Niayes

area, Saint-Louis (coef=coefﬁcient).

Month Decade Stage

coef.

Etc

mm/day

Etc

mm/decade

Oct 3 Init 0.60 3.11 2.8

Nov 1 Init 0.60 3.20 0.2

Nov 2 Init 0.60 3.28 0.0

Nov 3 Deve 0.66 3.55 0.0

Dec 1 Deve 0.81 4.34 0.2

Dec 2 Deve 0.96 5.12 0.2

Dec 3 Deve 1.13 6.20 0.3

Jan 1 Mid 1.22 6.95 0.4

Jan 2 Mid 1.22 7.19 0.5

Jan 3 Mid 1.22 7.45 0.5

Fev 1 Mid 1.22 7.71 0.6

Fev 2 Late 1.20 7.86 0.7

Fev 3 Late 1.11 7.55 0.5

Mar 1 Late 1.01 7.18 0.2

Mar 2 Late 0.92 6.77 0.0

ments will be sent to the system. These data will then

be combined with evapotranspiration data of the ac-

tual area to serve as an input to our system.

During data gathering and modelization (Figure

3), the system collects raw data from various sources.

The sensor network will measure the water kept in the

soil before irrigation (Wb). The modelization part,

raw data are annotated then integrated. We will use

CROPWAT’s generated data tables (Table 1) to have

the crop evapotranspiration value (ET

). These data

are organized in a model to become a state context of

a section of the ﬁeld. The processing phase will

receive the integrated data from the precedent stage.

During this phase, reasoning is applied on these inte-

grated data to infer an efﬁcient irrigation schedule for

each section of the ﬁeld. For the reasoning, a rule-

based engine can be utilized.

SMARTGREENS 2020 - 9th International Conference on Smart Cities and Green ICT Systems

172

Figure 3: Functioning architecture.

5 RESULT

We develop an adaptive and a contextual system to

automate and optimize irrigation. Based on that, an

adapted water supply is applied through the irriga-

tion system. As a result, the watering system will use

strict water quantity to allow the crop to grow prop-

erly. This output will be used to make a set of infer-

ence rules to adjust the water supply and optimize it.

In the processing phase, the output of the mod-

elization phase is used to create inference rules for

a more effective irrigation that will help predict wa-

tering plan when there is a fault in our system. The

system monitors the intensity of the rain, the over-

ﬂow due to a malfunction of the drip setup, to stop

the watering pump.

The sensor location in soil will be determinated by

the plant root zone. The Management Allowed Deple-

tion or MAD is the highest amount of water in plant

root zone that can be taken of. This amount depends

on many factors like the plant type, soil properties or

exposition of the land toward the sun wich is more

surfaces for evaporation. The value is estimated be-

tween 50% to 70% in plant root zone. Placing our

sensor in that point can contribute in optimization of

our irrigation plan Figure 4.

Our proposition is comparative to a real-time sys-

tem. In fact, the water need depend on actual state

of the soil. Rain, overﬂood, dysfunctioning sprinkler

or a continuous watering is automatically detected by

sensors and relayed to the system. Each sensor can be

located and if any issue, an action is made to cut off

actuators in this area.

Figure 4: Onion bulb root system.

6 CONCLUSIONS

Gross irrigation in Niayes area affects water resources

a lot. In addition of climatic changes, rainy seasons

are short. That causes water depletion in this area.

In this paper, we tried to show techniques used for

smart irrigation. We have looked at some examples

that uses mathematical models, application programs

and sensing devices to determine water use and water

need of crops.

Technology has contributed a lot in solutions of

water scarcity problem. In agriculture irrigation sys-

tems are progressively controlled by technology. To

monitor water use, devices like soil moisture sensors

help a lot in that point. Data gathered are integrated

to help make decision and adjust watering plans.

We presented a proposition of smart-irrigation.

Based on using sensors and exploiting advantages

given by FAO Penman-Monteith formula to get cer-

tain data like plant evapotranspiration. In the future

of our work we’ll compare our system with an another

that ties up to our work.

REFERENCES

Weather data packages.

Angus, I. and Butler, S. (2014). Une plan

ete trop peupl

ee?.

Le mythe populationniste, l’immigration et la crise

ecologique.

Ecosoci

e (

Editions).

Ba, A. and Cantoreggi, N. L. (2018). Agriculture urbaine

et p

eriurbaine (aup) et

economie des m

enages agri-

urbains

a dakar (s

egal). International Journal of En-

vironment, Agriculture and Biotechnology, 3(1):195–

207.

Approach using the Internet of Things in the Sahel for Smart Irrigation

173

Bricas, N. and Seck, P. A. (2004). L’alimentation des villes

du sud: les raisons de craindre et d’esp

erer. Cahiers

Agricultures, 13(1):10–14.

Chen, Y., Marek, G. W., Marek, T., Brauer, D. K., and

Srinivasan, R. (2018). Improving swat auto-irrigation

functions for simulating agricultural irrigation man-

agement using long-term lysimeter ﬁeld data. Envi-

ronmental Modelling & Software, 99:25–38.

George, B. A. (2017). Technical manual for ”crop water

requirements and irrigation scheduling”.

Mohanraj, I., Gokul, V., Ezhilarasie, R., and Uma-

makeswari, A. (2017). Intelligent drip irrigation and

fertigation using wireless sensor networks. In Techno-

logical Innovations in ICT for Agriculture and Rural

Development (TIAR), 2017 IEEE, pages 36–41. IEEE.

Myers, T., Mohring, K., and Andersen, T. (2017). Se-

mantic iot: Intelligent water management for efﬁcient

urban outdoor water conservation. In Joint Interna-

tional Semantic Technology Conference, pages 304–

317. Springer.

Niang, S. (1996). Utilisation des eaux us

ees domestiques en

mara

ıchage p

eriurbain

a dakar (s

egal). Science et

changements plan

etaires/S

echeresse, 7(3):217–223.

OpenWeatherMap.org. Weather api.

Savva, A. P. and Frenken, K. (2002). Crop water require-

ments and irrigation scheduling. FAO Sub-Regional

Ofﬁce for East and Southern Africa Harare.

SMARTGREENS 2020 - 9th International Conference on Smart Cities and Green ICT Systems

174