Tori.dvi

Size: px

Start display at page:

Download "Tori.dvi"

しょうりさわまつ
5 years ago
Views:

1 Analysis of Bird Count Series by Local Regression to Explore Environmental Changes Hideyasu SHIMADZU and Ritei SHIBATA Five bird count series observed on a monthly basis from 1967 to 1998 at Jiyu- Gakuen, Higashi-Kurume in Tokyo are simultaneously analysed. Each count series is decomposed into three components, long trend, short trend and irregular by two step loess smoothings. This decomposition explains well the relationship between the bird count and some of environmental changes. By selecting appropriate locations and scales as well as the smoothing parameters so as to minimise the residual sum of squares, it is shown that each five long trend very similarly moves with one of environmental factors. Turtledove (Streptopelia orientalis), Browneared Bulbul (Hypsipetes amaurotis) and Great Tit (Parus major) increased its number to link with the enlargement of housing area. Tree Sparrow (Passer montanus) and Gray Starling (Sturnus cineraceus) gradually decreased its number to link with the decrease of open field. Such two bird groups are well described by the environmental preferences. Variation of each short trend can be explained by the effects of breeding season or winter wandering. The fact that each irregular series has no significant trend and very low correlation coefficients suggests a success of our decomposition

2 Maeda , Bird Census 2

3 BTO British Trust for Ornithology 1962 Common Bird Census(CBC USGS the U.S. Geological Survey CWS the Canadian Wildlife Service 1966 Breeding Bird Survey(BBS km 6.5km 3.5km 12.92km 2 70m 40m 56m 47m 10 m 2 Acer palmatum Zelkova serrata Ginkgo biloba Carpinus tschonoskii Pinus densiflora Quercus myrsinaefolia Aucuba japonicapleioblastus chino 3

4 3.2 1: Territory Mapping Line Transect Point Count (Bibby et al. 2000) CBC BBS 4

5 (2002) DandD(Data and Description) URL A

6 1: Passer montanus Streptopelia orientalis Parus major Hypsipetes amaurotis Sturnus cineraceus Cyanopica cyana Carduelis sinica Corvus macrorhynchos Hirundo rustica Turdus naumanni eunomus Emberiza spodocephala Lanius bucephalus Bambusicola thoracica Motacilla alba Dendrocopos kizuki Cettia diphone Anas poecilorhyncha Coccothraustes coccothraustes Zosterops japonicus Eophona personata Motacilla cinerea Columba livia var. domestica Corvus corone Motacilla grandis Phoenicurus auroreus Picus awokera Lanius tigrinus Accipiter gularis Egretta garzetta Emberiza cioides Emberiza rustica Parus ater 33 9 Cuculus canorus 34 8 Delichon urbica 35 8 Turdus pallidus 36 7 Cuculus saturatus 37 7 Ninox scutulata 6

7 38 6 Ficedula narcissina 39 6 Garrulus glandarius 40 6 Phylloscopus coronatus 41 4 Accipiter gentilis 42 4 Alauda arvensis 43 3 Psittacula krameri manillensis 44 2 Apus pacificus 45 2 Milvus migrans 46 2 Parus varius 47 2 Pericrocotus divaricatus 48 2 Tarsiger cyanurus 49 2 Turdus chrysolaus 50 2 Turdus naumanni naumanni 51 1 Anas crecca 52 1 Anthus hodgsoni 53 1 Caprimulgus indicus 54 1 Cyanoptila cyanomelana 55 1 Ficedula parva 56 1 Larus ridibundus 57 1 Muscicapa griseisticta 58 1 Nycticorax nycticorax 59 1 Pernis apivorus 60 1 Zoothera dauma

8 : 4 1 8

9 0 4.1 lowess(cleveland 1979) loess(cleveland and Devlin 1988) 1994 Fan and Gijbels (1996) James et al. (1996) BBS American Warblers (t j,y j ),j =1,...,n f (t) f (t) p w ( ) n ( ) tj t w {y j f t (t j )} 2 min (1) d δ (t) ft j=1 f t d δ (t) = max j;tj U δ (t) t j t U δ (t) t [nδ] [nδ] nδ δ span n f t f t (s) = p β k (t)(s t) k k=0 9

10 p S-PLUS loess w 3 3 { (1 x 3 ) 3, 0 x<1 w (x) = 0, loess degree p ˆβ 0 (t) p =0 ˆβ 0 (t) = j 1 s 0 (t) w ( ) tj t y j d δ (t) p =1 ˆβ 0 (t) = j s 2 (t) (t j t) s 1 (t) s 0 (t) s 2 (t) s 1 (t) 2 w ( ) tj t y j d δ (t) p =2 ˆβ 0 (t) = 1 c (t) [ {s2 (t) s 4 (t) s 3 (t) 2} (t j t) {s 1 (t) s 4 (t) s 2 (t) s 3 (t)} j +(t j t) 2 { s 1 (t) s 3 (t) s 2 (t) 2}] w ( ) tj t y j d δ (t) s r (t) = j (t j t) r w ( t j t /d δ (t)), c(t) =2s 1 (t) s 2 (t) s 3 (t) + s 0 (t) {s 2 (t) s 4 (t) s 3 (t)} s 1 (t) 2 s 4 (t) s 2 (t) 3 y 1,...,y n [nδ] t = t j p =1 p =0 p 4.2 δ span2 10

11 日本統計学会誌 ( 第 34 巻シリーズ J 第 2 号 pp ) original long.trend short.trend irregular : 1 2 B Shibata and Miura(1997) 3 S-PLUS x t =1,..., i 3 Z i (t) =L i (t)+s i (t)+i i (t),t=1,...,384. L i (t) S i (t) 11

12 L i (t) S i (t) 1 L i (t),s i (t) I i (t) L i (t) S i (t) I i (t) 3 S-PLUS loess δ span p degree Fan and Gijbels Z i (t) 2 4 4, km 2 km 2 12

13 km km 4: 4 R (t) b i a i a i,b i 384 t=1 {R (t) (a i L i (t)+b i )} 2 ai,b i min (2) (2) a i,b i L i (t) δ δ (2) δ a i,b i δ a i,b i (2) δ a i,b i 13

14 2: span p =1 2 nδ?? a i,b i 3 2?? a i,b i ,5 (2) δ a i,b i p =1 2 6,

15 km Year 5: 5 3: 15

16 km Year 6: 5 5 L i (t) 5.2 Z i (t) L i (t) Z i (t) L i (t) p =1 p =2 7 p =1 p =

17 :

18 : 18

19 Jan Feb Mar Apl May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apl May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apl May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apl May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apl May Jun Jul Aug Sep Oct Nov Dec 9:

20 : 4:

21 I i (t)

22 : 22

23 A DandD (2002) DandD(Data and Description) URL DandD DandDR R R R > library(dad) dad > bird <- DandD("JiyuBirdCount.dad") DandD > plot(suzume) B 2 S- PLUS S R Ver R > library(modreg) 23

24 decomp <- function(original, span1, span2, plot = F) { # # original: original time series # span1: long span (year) # span2: short span (year) # plot: graph plot # loess1 <- loess(original ~ c(1:length(original)), span = (span1 * 12)/length(original), degree = 1, family = "symmetric") long.trend <- loess1$fit tsp(long.trend) <- tsp(original) long.irregular <- loess1$residuals loess2 <- loess(long.irregular ~ c(1:length(original)), span = (span2 * 12)/length(original), degree = 2, family = "symmetric") short.trend <- loess2$fit tsp(short.trend) <- tsp(original) irregular <- loess2$residuals tsp(irregular) <- tsp(original) decomp <- list(original = original, long.trend = long.trend, short.trend = short.trend, irregular = irregular) if(plot == T) { par(mfrow = c(4, 1), lab = c(38, 10, 7)) plot(decomp$original, type = "l", ylab = "", xlab = "") title("original") plot(decomp$long.trend, type = "l", ylab = "", xlab = "") title("long.trend") plot(decomp$short.trend, type = "l", ylab = "", xlab = "") title("short.trend") plot(decomp$irregular, type = "h", ylab = "", xlab = "") title("irregular") } return(decomp) } [1] 1984 Strix 3:

25 [2] Colin J. Bibby, Neil D. Burgess, David A. Hill and Simon H. Mustoe Bird Census Techniques 2nd. ed. Academic Press, London. [3] William S. Cleveland Robust Locally Weighted Regression and Smoothing Scatterplots. Journal of the American Statistical Association 74(368): [4] William S. Cleveland and Susan J. Devlin Locally Weighted Regression: An Approach to Regression Analysis by Local Fitting. Journal of the American Statistical Association 83(403): [5] 1994 S [6] J. Fan and I. Gijbels Local Polynomial Regression and Its Applications. Chapman&Hall, London. [7] 1982 Strix 1: [8] 1985 Strix 4:33-42 [9] 1989 Strix 8: [10] Strix 14: [11] Frances C. James, Charles E. McCulloch and David A. Wiedenfeld New Approaches to The Analysis of Population Trends in Land Birds. Ecology 77(1): [12] 2000 [13] : [14] : [15] : [16] 1994 Strix 13:

26 [17] Taku Maeda Bird Communities and Habitat Relationships in a Residential Area of Tokyo. Journal of the Yamashina Institute for Ornithology 30: [18] 1988 Strix 7: [19] 1996Strix 14: [20] Ritei Shibata and Ryozo Miura Decomposition of Japanese Yen Interest Rate Data Through Local Regression. Financial Engineering and the Japanese Markets 4: [21] 2001 [22] [23] : [24] 2003 Strix 21: [25] 2001 DandD 49:

佐藤.indd

佐藤.indd (29): 73-82 Mar. 2007 Yuya Sato, Yasuhiro Ishikawa, Shinpei Seki, Yuki Yoshida, Kouichiro Baba and Masaaki Fujiyoshi : Wild Birds Observed on the Shonan Campus of Tokai University 50ha 50 50 1992 50 20m