如何解决R中sdm包创建sdmData对象的报错问题？

问题场景 #

使用R语言的sdm包构建物种分布模型，运行以下代码：

sdmData <- sdmData(sp~ Wetland +Distance_Protected_Area +Distance_Shore + 
                    Distance_Settlement +Distance_Cages + 
                    Cages_Present + Bathymetry,
  train = sdm_Data, predictors = AllOccRasters, bg = BackgroundPoints)

第一次报错 #

运行后出现如下错误：

Error in (function (classes, fdef, mtable)  : 
  unable to find an inherited method for function ‘sdmData’ for signature ‘"formula", "data.frame", "RasterStack"’

尝试的解决方法及新报错 #

尝试将数据框转换为空间对象：

sdm_Data <- as(sdm_Data, 'Spatial')

但出现新错误：

Error in as(sdm_Data_nets, "Spatial") : 
  no method or default for coercing “data.frame” to “Spatial”

数据情况 #

sdm_Data是经过na.omit()处理后的data.frame，包含物种列（sp）、环境变量列，以及latitude和longitude坐标列，数据结构示例：

structure(list(sp = c("F_ardesiaca", "F_ardesiaca", "F_ardesiaca", "F_ardesiaca", "F_ardesiaca"), 
    Totora = c(-0.239973815, -0.239973815, -0.239973815, 
    -0.239973815, -0.239973815), Distance_Protected_Area = c(-1.378201598, 
    -1.123927933, -1.137970591, -1.123927933, -1.162127222), 
    Distance_Shore = c(-1.094753612, -0.960762706, -0.911857012, 
    -0.960762706, -1.086927111), Distance_Settlement = c(-1.100245165, 
    -0.972346735, -0.935389232, -0.972346735, -1.141459148), 
    Distance_Cages = c(-1.445217077, -1.437232488, -1.260983258, 
    -1.437232488, -1.181728985), Cages_Present = c(0L, 0L, 
    0L, 0L, 0L), Bathymetry = c(1.122127163, 0.740096285, 
    0.369610894, 0.740096285, 0.836073376), latitude = c(-15.86403, 
    -15.48954, -15.50397, -15.4846, -15.5163), longitude = c(-69.93177, 
    -69.88749, -69.8816, -69.89034, -69.87735)), na.action = structure(c(`1` = 1L, 
`2` = 2L, `46` = 46L, `71` = 71L, `73` = 73L, `76` = 76L, `77` = 77L, 
`87` = 87L, `89` = 89L, `113` = 113L, `140` = 140L, `141` = 141L, 
`142` = 142L, `143` = 143L, `144` = 144L, `147` = 147L, `152` = 152L, 
`163` = 163L, `174` = 174L, `185` = 185L, `188` = 188L, `193` = 193L, 
`196` = 196L, `207` = 207L, `218` = 218L, `229` = 229L, `240` = 240L, 
`241` = 241L, `252` = 252L, `263` = 263L, `274` = 274L, `285` = 285L, 
`296` = 296L, `307` = 307L, `318` = 318L, `327` = 327L, `328` = 328L, 
`329` = 329L, `340` = 340L, `351` = 351L, `352` = 352L, `363` = 363L, 
`374` = 374L, `385` = 385L, `396` = 396L, `407` = 407L, `418` = 418L, 
`429` = 429L, `440` = 440L, `451` = 451L, `462` = 462L, `463` = 463L, 
`467` = 467L, `468` = 468L, `474` = 474L, `485` = 485L, `493` = 493L, 
`496` = 496L, `507` = 507L, `516` = 516L, `518` = 518L, `524` = 524L, 
`525` = 525L, `526` = 526L, `529` = 529L, `540` = 540L, `551` = 551L, 
`554` = 554L, `562` = 562L, `566` = 566L, `573` = 573L, `574` = 574L, 
`585` = 585L, `596` = 596L, `607` = 607L, `618` = 618L, `629` = 629L, 
`640` = 640L, `651` = 651L, `662` = 662L, `673` = 673L, `682` = 682L, 
`684` = 684L, `685` = 685L, `687` = 687L, `688` = 688L, `690` = 690L, 
`693` = 693L, `694` = 694L, `696` = 696L, `707` = 707L, `718` = 718L, 
`729` = 729L, `740` = 740L, `751` = 751L, `762` = 762L, `773` = 773L, 
`782` = 782L, `784` = 784L, `795` = 795L, `796` = 796L, `807` = 807L, 
`808` = 808L, `818` = 818L, `829` = 829L, `840` = 840L, `851` = 851L, 
`862` = 862L, `873` = 873L, `874` = 874L, `878` = 878L, `884` = 884L, 
`885` = 885L, `888` = 888L, `895` = 895L, `903` = 903L, `906` = 906L, 
`907` = 907L, `918` = 918L, `921` = 921L, `929` = 929L, `934` = 934L, 
`935` = 935L, `937` = 937L, `939` = 939L, `940` = 940L, `948` = 948L, 
`950` = 950L, `951` = 951L, `952` = 952L, `954` = 954L, `956` = 956L, 
`962` = 962L, `969` = 969L, `973` = 973L, `974` = 974L, `984` = 984L, 
`992` = 992L, `995` = 995L, `998` = 998L, `999` = 999L, `1004` = 1004L, 
`1006` = 1006L, `1017` = 1017L, `1018` = 1018L, `1029` = 1029L, 
`1053` = 1053L, `1054` = 1054L, `1073` = 1073L, `1084` = 1084L, 
`1095` = 1095L, `1106` = 1106L), class = "omit"), row.names = 3:7, class = "data.frame")

问题分析 #

1. 第一个报错的核心原因：sdmData函数要求train参数必须是空间数据对象（如SpatialPointsDataFrame），而不是普通的data.frame。
2. 直接用as(sdm_Data, 'Spatial')失败，是因为R无法直接将带坐标列的data.frame自动识别为空间对象，必须显式指定坐标信息和坐标参考系（CRS）。

解决方案 #

需要将sdm_Data转换为SpatialPointsDataFrame，步骤如下：

1. 加载sp包（处理空间数据的基础包）；
2. 提取坐标列，创建空间点对象并指定正确的CRS（需与predictors参数中的栅格数据CRS一致）；
3. 组合属性数据与空间点，生成SpatialPointsDataFrame。

示例代码：

# 加载sp包
library(sp)

# 提取longitude和latitude作为坐标
coords <- sdm_Data[, c("longitude", "latitude")]

# 创建SpatialPoints对象，这里以WGS84（EPSG:4326）为例，需根据你的栅格数据CRS调整
proj_crs <- CRS("+proj=longlat +datum=WGS84 +no_defs")
sp_points <- SpatialPoints(coords, proj4string = proj_crs)

# 转换为SpatialPointsDataFrame，保留原数据的属性列（移除重复的坐标列）
sdm_Data_sp <- SpatialPointsDataFrame(sp_points, data = sdm_Data[, !names(sdm_Data) %in% c("longitude", "latitude")])

# 再次运行sdmData函数
sdmData <- sdmData(sp~ Wetland +Distance_Protected_Area +Distance_Shore + 
                    Distance_Settlement +Distance_Cages + 
                    Cages_Present + Bathymetry,
  train = sdm_Data_sp, predictors = AllOccRasters, bg = BackgroundPoints)

注意事项 #

• 确保proj_crs与AllOccRasters的CRS完全一致，可通过crs(AllOccRasters)查看栅格的CRS，然后复制对应的参数；
• 如果CRS不一致，需用spTransform(sdm_Data_sp, crs(AllOccRasters))转换空间对象的CRS。

平台选择建议 #

这个问题同时涉及R语言代码调试和空间数据处理：

• 若侧重包函数使用、代码报错排查，Stack Overflow更合适；
• 若侧重空间数据GIS层面的处理逻辑，GIS Stack Exchange也适配。
  当前问题核心是数据类型转换和函数参数要求，Stack Overflow的匹配度更高。

内容的提问来源于stack exchange，提问作者Birdman