Runoff generation and

sediment transport: Do saturated zones play a role in

tropical watersheds?

CHRISTIAN DAVID GUZMÁN

PHD CANDIDATE, CORNELL UNIVERSITY

FOOD SYSTEMS AND POVERTY REDUCTION IGER FELLOW

NSF/USAID RESEARCH AND INNOVATION FELLOW

26 NOVEMBER 2014

Motivation

Soil and water conservation is a multi-actor

unresolved solution

Heavy sediment loads reveals symptoms

Sustainability requires a focus on processes in

the Andean climate, specifically:

1. Main underlying flow patterns

2. Sediment concentration patterns and

variability

2

CALI, COLOMBIA

Infiltration

Evaporation

Soil moistureInterflow

Evapotranspiration

Recharge

Precipitation

Runoff

Runoff

Water table Evapotranspiration

Rainfall intensity has to

exceed the soil’s infiltration

rate for Hortonian flow

4

0

2

4

6

8

10

0.00 0.20 0.40 0.60 0.80 1.00

Infiltra

tio

n c

ap

ac

ity

,

Ra

infa

ll In

ten

sity

(c

m/h

r)

Probability of Exceedance

Median infiltration rate

Minimum infiltration rate

Rainfall intensity curve

Bayabil et al., 2010

Mecanism:

-Hortonian(Horton, 1933)

-Dunnean

(Dunne and Black, 1970,

Kirkby, 1969

Freeze, 1972

Dunne et al., 1975)

MAYBAR, ETHIOPIA

5Sediment concentration data

exhibit variable but particular

underlying patterns

Inst

alla

tio

n o

f te

rra

ce

s

Cw = AcacPce• Rd 0.4

Weirs

Terraces

C = a• Q n

Guzman et al., 2013; Steenhuis et al., 2014 ANJENI, ETHIOPIA

Runoff generation and sediment transport dependent on storage

capacity

Hillslope

Degraded

Saturated

infiltration

interflow

overland flow

(Cappus, 1960,Kirkby, 1969; Freeze, 1972;Dunne and Black, 1979; Beven, 2000; Buytaert et al., 2007; Collick et al., 2009; Steenhuis et al., 2013; Tilahun et al., 2014)

ANDIT TID, ETHIOPIA

Objectives

Short term: detect runoff and sediment

transport patterns in Aguaclara watershed

network

Mid term: build supporting evidence for a

well defined runoff mechanism hypothesis

Long term: study hydrological and

geomorphological patterns in the Andes for

conservation adjustment

7

COLOMBIA

Objectives

How do spatial and temporal

patterns reveal which dynamics are

present?

8

Middle

Upper

Lower

R² = 0.6235

0

2

4

6

8

10

12

14

25

-Ju

n

15

-Ju

l

4-A

ug

24

-Au

g

13

-Se

p

Se

dim

en

t C

on

ce

ntr

atio

n (

g L

-1)

Mean daily sediment

concentration

D.C. Dagnew, 2013DEBRE MAWI, ETHIOPIA

COLOMBIA

Questions

1. Which runoff generation mechanism is

dominantly present in a representative

watershed?

2. Which areas of the site might be sediment

source areas?

3. Is there a correlation between soil loss and

nutrient loss?

4. Which hydrological and pedological dynamics can be adjusted for improved runoff and

sediment transport estimation?

9

CALI, COLOMBIA

Methods 10

Rio Bolo

Micro-cuenca

La Vega

The Nature Conservancy

Fondo Agua por La Vida y La Sostenibilidad, Asobolo, Asocaña, Cenicaña

AGUACLARA, COLOMBIA

Methods

1. Characterization of flow patterns and sediment

and nutrient export

Rainfall intensity vs infiltration capacity (Double ring

infiltrometer, Constant head permeameter)

Monitoring hydrological balance

Monitoring sediments and water

2. Spatial and temporal changes in the micro

watershed

Soil surface changes

Water table depths

11

AGUACLARA, COLOMBIA

DEBRE MAWI, ETHIOPIA

Methods

3. Soil nutrient status and relation to soil

loss patterns (0-15 cm)

Macronutrients (nitrogen, potassium, phosphorus)

Exchangeable Cations

pH, organic matter

4. Comparison of patterns

PED, TOPMODEL

InVEST, RIOS

SWAT

12

AGUACLARA, COLOMBIA

Analysis: Broad scale

1. Conceptualization of hydrology

Flow reservoir transfer

(Thornthwaite-Mather, 1955)

Nash-Sutcliffe Efficiency

Conceptualization of sediment transport

Stratification of data (cumulative effective precipitation; Lui et al., 2008)

Non-parametric statics, ej. Kruskal-Wallis, Wilcoxon Rank Sum

13

AGUACLARA, COLOMBIA

14Analysis: Broad scale

Tilahun et al., 2014 AGUACLARA, COLOMBIA

Analysis: Localized scale

2. Generate localized patterns of soil loss on hillslope and land use areas

Upslope vs downslope, etc.

Grazing vs forests, etc.

3. Compare nutrient status and change patterns with soil depth changes

Coefficient of determination, R2

Correlation coefficient, Pearson r

15

AGUACLARA, COLOMBIA

16

-3.20

-2.70

-2.20

-1.70

-1.20

-0.70

-0.20

25-Jul 4-Aug 14-Aug 24-Aug 3-Sep

Wat

er t

able

dep

th b

elow

surf

ace

(m)

Downslope w/o Gully

Midslope-1

Midslope-2

Upslope

8

9

16

1

Midslope-1

Midslope-2

Upslope

Downslope Weir

Analysis: Coupling scales

1. Conceptualization of hydrology

2. Generate localized patterns of soil loss on hillslope and land use areas

DEBRE MAWI, ETHIOPIA

Analysis: Critical concepts

4. Comparison of pattern representation with modeling NSE , RMSE, GLUE

17

AGUACLARA, COLOMBIA

Expected outcomes

1. Preliminary integration of runoff concepts

2. Identify areas or land uses that generate runoff and sediment

3.Identify the simple or complex relationship between soil loss and nutrient changes

4. Evaluation of model performance*

18

Horton, 1933; Dunne and Black, 1979 AGUACLARA, COLOMBIA

Future work

Study ecosystem services in greater

detail

Develop modeling alternatives

Replicate study

19

AGUACLARA, COLOMBIA

Thank you!

cdg65@cornell.edu

20

ETHIOPIACOLOMBIA

Montgomery, 2007