Predicting the winners and losers of agricultural change · Predicting the winners and losers of...

Preview:

Citation preview

Predicting the winners and losersof agricultural change

Jonathan StorkeyRothamsted Research

Predicting winners and losers ofagricultural change

• The big picture for arable weeds post war

• Looking back; understanding past changes in weed floras

• Looking forward; predicting the impact of future change

Post war changes in crop husbandry and effectson yields

+Landscape scale changes

Simpler rotationsLarger fieldsBlock-cropping

Agricultural landscapes have become more homogeneous both in space and time

Common trends in arable floras

Common trends in arable floras

• Weed diversity and abundance at the Relevé scale (α ) dramatically declines

Largely an effect of herbicides…raising the bar for the winners

• Shallower declines in regional diversity (γ diversity)

Refuges or alternative habitats are important for maintaining populations

• Neophytes and generalists have increased at the expense of archaeophytesand specialists (+abundance based mechanisms)

Homogeneity of landscapes eg. > N inputs

• β diversity increasingly important for maintaining in-field weed diversity

Population change index

1962 Atlas surveyed 10km squaresover Britain and Ireland for speciespresence / absence between 1955and 1960.

2000 Atlas repeated survey between1987 and 1999.

Proportion of squares in which eachspecies was recorded calculated foreach period and logit transformed.

Linear regression with earlier periodas explanatory variable.

Change index calculated as thestandardised residual for each species.

Population change index = -2.19

Examples of New Atlas data:Adonis annua

Preston, C.D., Pearman, D.A. & Dines, T.D. (2002) New Atlas of the British and Irish Flora Oxford University Press, Oxford, UK.

Population change index = -3.65

Examples of New Atlas data:Scandix pecten-veneris

Population change index = -4.78

Examples of New Atlas data:Galium tricornutum

Arable specialists have declined disproportionately

F pr. = 0.001

0

10

20

30

40

50

60

70

80

90

100

Population change index

Freq

uenc

y

-5 -4 -3 -2 -1 0 1 2 3

Corn cleavers Black-grass

Arable specialists have declined disproportionately

Which specialists are most vulnerable on aEuropean scale?

European analysis of rare and threatened weeds

• A questionnaire was sent to agricultural botanists in 29 European countries.

• They were asked to identify weeds on the National Red data list and supplementedwith expert knowledge on species they know are threatened or declining nationally.

• They were also asked to identify drivers of change.

• These data were combined with data on trends in agro-chemical use and landscapevariables for each country.

• GLMs and redundancy analysis were used to analyse trends.

European analysis of rare and threatened weeds:Response variables

European analysis of rare and threatened weeds:Explanatory variables

European analysis of rare and threatened weeds:Results

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

0 2 4 6 8 10

Prop

ortio

n ra

re o

r thr

eate

ned

arab

le p

lant

s

2008 wheat yield (t/ha)

European analysis of rare and threatened weeds:Results

0.41 - 0.6

0 - 0.2

0.61 - 0.8

0.21 - 0.4

no available data

European analysis of rare and threatened weeds:Results

European analysis of rare and threatened weeds:Results

0

5

10

15

20

25

Car

yoph

ylla

ceae

Ast

erac

eae

Bra

ssic

acea

ePo

acea

eSc

roph

ular

iace

aeA

piac

eae

Ran

uncu

lace

aeLa

mia

ceae

Papa

vera

ceae

Bor

agin

acea

eLi

liace

aePr

imul

acea

eC

heno

podi

acea

eR

ubia

ceae

Val

eria

nace

aeEu

phor

biac

eae

Faba

ceae

Mal

vace

aeG

eran

iace

aeC

ampa

nule

acea

eR

osac

eae

Lyth

race

aeC

onvo

lvul

acea

eH

yper

icac

eae

Irid

acea

eJu

ncac

eae

Lina

ceae

Oro

banc

hace

aeSo

lana

ceae

Thym

elae

acea

e

Num

ber o

f rar

e or

thre

aten

ed sp

ecie

s **

*

*

*

*

***

****

European analysis of rare and threatened weeds:Results

Can we model the filtering effect of agriculturalchange on weed floras (still looking back)?

Booth BD & Swanton CJ. (2002) Assembly theory applied to weed communities. Weed Science 50: 2-13

Can we model the filtering effect of agriculturalchange on weed floras (still looking back)?

The Ecological Matrix Approach

Habitat Germination Flowering Ellenberg Numbers

Family / species Common name Hedge Verge Field spring autumn 4 5 6 7 Nitrogen Moisture Affinity Change index

AraceaeArum maculatum Cuckoo Pint 1 1 1 1 1 7 5 3 -0.28

BoraginaceaeAnchusa arvensis Bugloss 1 1 1 1 1 5 4 1 -0.7Cynoglossum officinale Hound's-tongue 1 1 1 1 1 6 4 2 -1.09Echium vulgare Viper's Bugloss 1 1 1 1 1 1 4 4 2 -0.24Lithospermum arvense Field Gromwell 1 1 1 1 1 1 5 4 1 -1.91Myosotis arvensis Field Forget-me-not 1 1 1 1 1 1 1 1 1 6 5 2 -0.34Myosotis ramosissima Early Forget-me-not 1 1 1 1 1 3 3 1 0.11

CaprifoliaceaeLonicera periclymenum Honeysuckle 1 1 1 1 5 6 3 -0.11

CaryophyllaceaeCerastium fontanum Mouse-ear Chickweed 1 1 1 1 1 1 1 1 4 5 3 1.4Silene dioica Red Campion 1 1 1 1 1 1 1 7 6 3 -0.44Silene gallica Small-flowered catchfly 1 1 1 1 1 5 4 1 -2.78Silene latifolia White Campion 1 1 1 1 1 1 6 4 2 -0.88Silene noctiflora Night-flowering catchfly 1 1 1 1 6 4 1 -2.04Silene vulgaris Bladder Campion 1 1 1 1 1 1 5 4 2 -1.26Spergula arvensis Corn Spurrey 1 1 1 1 1 5 4 1 -2.3Spergularia rubra Sand Spurrey 1 1 1 1 1 2 3 2 0.05Stellaria media Chickweed 1 1 1 1 1 1 1 7 5 2 0.03

ChenopodiaceaeAtriplex patula Orache 1 1 1 1 7 5 2 -0.34Atriplex prostrata Hastate Orache 1 1 1 7 7 2 1.1Chenopodium album Fat-hen 1 1 1 7 5 1 -0.73Chenopodium ficifolium Fig-leaved Goosefoot 1 1 1 1 7 6 2 1.9Chenopodium polyspermum Many-seeded Goosefoot 1 1 1 8 6 1 0.62

Autumn drilling(No spring germination in crop)

Earlier harvest(Reduced July flowering in crop)

Family / Species Habitat filter Germination Score

Habitat filter Germination Flowering Score

AraceaeArum maculatum 0 1 0 0 0 0 0.00

BoraginaceaeAnchusa arvensis 1 0 0 1 2 2 0.25Cynoglossum officinale 0 1 0 0 0 3 0.00Echium vulgare 2 0 0 2 2 2 0.13Lithospermum arvense 1 0 0 1 2 3 0.17Myosotis arvensis 3 0 0 3 2 3 0.06Myosotis ramosissima 0 -1 0 0 1 0 0.00

CaprifoliaceaeLonicera periclymenum 0 1 0 0 0 2 0.00

CaryophyllaceaeCerastium fontanum 2 0 0 2 2 4 0.06Silene dioica 0 0 0 0 2 4 0.00Silene gallica 1 0 0 1 2 2 0.25Silene latifolia 0 1 0 0 0 3 0.00Silene noctiflora 1 0 0 1 2 1 0.50Silene vulgaris 0 1 0 0 0 2 0.00Spergula arvensis 1 0 0 1 2 2 0.25Spergularia rubra 1 0 0 1 2 2 0.25Stellaria media 1 0 0 1 2 4 0.13

ChenopodiaceaeAtriplex patula 1 1 1 1 0 2 0.00Atriplex prostrata 1 1 1 1 0 1 0.00Chenopodium album 1 1 1 1 0 1 0.00Chenopodium ficifolium 1 0 0 1 2 1 0.50Chenopodium polyspermum 1 1 1 1 0 1 0.00

The Ecological Matrix Approach

The Ecological Matrix Approach

Analysis of weeds on Broadbalk (a gradientof fertiliser use); Ellenbergà traits

• Begun in 1843 to compare effect of inorganic fertilisers on wheat yield with farm yard manure.

• Originally, plots ran the length of the field (320 x 6m) but in 1926 they were split into sections which were sequentially fallowed to control weeds.

• Majority of experiment is still in continuous winter wheat with the exception of some plots which now have a wheat / oats / maize rotation

• Innovations in crop husbandry have been incorporated into the experiment including the use of herbicides since 1964.

• Section 8 has never received herbicide

High fertility plots

Analysis of weeds on Broadbalk (a gradientof fertiliser use)

Low fertility plots

Analysis of weeds on Broadbalk (a gradientof fertiliser use)

Analysis of weeds on Broadbalk (a gradientof fertiliser use)

Corn cleavers

Shepherd’s needle

9

10

11

12

13

14

15

16

0 kg/ha 48 kg/ha 96 kg/ha 144 kg/ha 192 kg/ha 240 kg/ha 288 kg/ha

Mea

n nu

mbe

r of s

peci

es re

cord

ed in

an

nual

sur

veys

(199

1-20

02)

Plot treatment

Moss, S.R., Storkey, J., Cussans, J.W., Perryman, S.A.M. & Hewitt, M.V. (2004) The Broadbalk long-term experiment at Rothamsted: what has it told us about weeds? Weed Science, 52(5), 864-73.

Broadbalk weeds – species richness

Broadbalk weeds – filtering effect of fertiliser?

*outstanding paper in Weed Science 2010 ☺

Broadbalk weeds – filtering effect of fertiliser?

Broadbalk weeds – filtering effect of fertiliser?

0

10

20

30

40

50

60

70

0 100 200 300Kg N ha-1

Num

ber o

f rec

ords

0

20

40

60

80

100

120

140

160

0 100 200 300Kg N ha-1

Num

ber o

f rec

ords

Winter annuals Rare weed trait syndrome(short, large seed, late flowering)

Broadbalk weeds – filtering effect of fertiliser?

Broadbalk weeds – filtering effect of fertiliser?

Height Seedmass

FloweringPopulationChange index

Predicting impact of future change inagricultural practice – looking forwards.

Modelling at the level of functional traits

R² = 0.41(P=0.035)

-2

-1

0

1

2

3

4

5

0 50 100 150 200 250Ln w

eeds

m-2

for 5

% y

ield

loss

Maximum height (cm)

Relationships between traits and model parameters

R² = 0.74(P<0.001)

0

1

2

3

4

5

6

7

8

-3 -2 -1 0 1 2 3y-

inte

rcep

t of L

n (s

eed

prod

uctio

n) o

n Ln

(sho

ot b

iom

ass

at m

atur

ty)

Ln seed weight (mg)

For the management scenario of high herbicide and fertiliser use, populations growth rate contours have been generated by the model for weeds with different combinations of height and seed weight and species

mapped onto the trait space using data from the database (�) rare weeds, (�) common weeds

Selection pressure towardstall and / or small seededweeds

Validation

Post-script: future improvement

Empirical data supporting trait-based approach

Are these my ‘false’ positives?

Thank you

Recommended