`mae.ensembleBMAgamma0` <-
function(fit, ensembleData, nSamples=NULL, seed=NULL, dates=NULL, ...) 
{
 weps <- 1.e-4

 M <- matchEnsembleMembers(fit,ensembleData)
 nForecasts <- ensembleSize(ensembleData)
 if (!all(M == 1:nForecasts)) ensembleData <- ensembleData[,M]

# remove instances missing all forecasts, obs, or dates

 M <- apply(ensembleForecasts(ensembleData), 1, function(z) all(is.na(z)))
 M <- M | is.na(ensembleVerifObs(ensembleData))
 M <- M | is.na(ensembleDates(ensembleData))
 ensembleData <- ensembleData[!M,]
 
 if (is.null(obs <- ensembleVerifObs(ensembleData)))
   stop("verification observations required")

#nObs <- length(obs)
 nObs <- ensembleNobs(ensembleData)

 if (!is.null(seed)) set.seed(seed)

 dateTable <- names(fit$nIter)

 if (!is.null(dates)) {

   dates <- sort(unique(as.character(dates)))

   if (length(dates) > length(dateTable)) 
     stop("parameters not available for some dates")

   K <- match(dates, dateTable, nomatch=0)

   if (any(!K) || !length(K)) 
     stop("parameters not available for some dates")

 }
 else {
   dateTable <- fit$dateTable
   K <- 1:length(dateTable)
  }

 dates <- names(dateTable)

## match dates in data with dateTable
 if (is.null(ensDates <- ensembleDates(ensembleData))) {
   if (length(dates) > 1) stop("date ambiguity")
   Dates <- rep( dates,nObs)
 }
 else {
   Dates <- as.character(ensDates)
   L <- as.logical(match( Dates, dates, nomatch=0))
   if (all(!L) || !length(L)) 
     stop("model fit dates incompatible with ensemble data")
   Dates <- Dates[L]
   ensembleData <- ensembleData[L,]
   obs <- obs[L]
   nObs <- length(obs)
 }

 Q <- as.vector(quantileForecast( fit, ensembleData, dates=dates))

 sampleMedian <- sampleMean <- predictiveMean <- rep(NA, nObs)
 names(sampleMedian) <- ensembleObsLabels(ensembleData)

 nForecasts <- ensembleSize(ensembleData) 
 ensembleData <- ensembleForecasts(ensembleData)

 l <- 0
 for (d in dates) {

    l <- l + 1
    k <- K[l]

    WEIGHTS <- fit$weights[,k]
    if (all(Wmiss <- is.na(WEIGHTS))) next

    I <- which(as.logical(match( Dates, d, nomatch = 0)))

    for (i in I) {
    
       f <- ensembleData[i,]

       M <- is.na(f) | Wmiss
     
       VAR <- fit$varCoefs[1,k] + fit$varCoefs[2,k]*f

       fTrans <- sapply(f, fit$transformation)

       MEAN <- apply(rbind(1, fTrans) * fit$biasCoefs[,,k], 2, sum)

       W <- WEIGHTS
       if (any(M)) {
         W <- W + weps
         W <- W[!M]/sum(W[!M])
       }

       predictiveMean[i] <- sum(W*MEAN[!M])

    if (!is.null(nSamples)) {
       RATE <- MEAN/VAR
       SHAPE <- MEAN*RATE

       PROB0 <- sapply(apply(rbind( 1, fTrans, f==0)*fit$prob0coefs[,,k],
                              2,sum), inverseLogit)
       RATE <- RATE[!M]
       SHAPE <- SHAPE[!M]
       PROB0 <- PROB0[!M]

       SAMPLES <- sample( (1:nForecasts)[!M], size = nSamples, 
                          replace = TRUE, prob = W)

       tab <- numeric(sum(!M))
       names(tab) <- (1:nForecasts)[!M]
       tabSamples <- table(SAMPLES)
       tab[names(tabSamples)] <- tabSamples

       Z <- tab == 0

       tab <- tab[!Z]

       SHAPE <- SHAPE[!Z]
       RATE <- RATE[!Z]
       PROB0 <- PROB0[!Z]

       I <- seq(along = tab)

       tab0 <- table(unlist(apply( cbind( seq(along = tab), tab), 1,
              function(nj,PROB0) {
         sample(c(nj[1],0), size = nj[2], replace = TRUE,
                prob = c(1-PROB0[nj[1]],PROB0[nj[1]]))}, 
                PROB0 = PROB0)))

       I <- as.numeric(names(tab0[-1]))
       tab[] <- 0
       tab[I] <- tab0[-1]

       Z <- tab == 0

       tab <- tab[!Z]

       if (length(tab)) {

          SHAPE <- SHAPE[!Z]
          RATE <- RATE[!Z]

          S <- apply( cbind( seq(along = tab), tab), 1,
              function(nj,SHAPE,RATE) 
                  rgamma(nj[2], shape=SHAPE[nj[1]], rate=RATE[nj[1]]),
                                      SHAPE = SHAPE, RATE = RATE)
           
# model is fit to the cube root of the forecast

         S <- sapply(as.vector(unlist(S)),
                           fit$inverseTransformation)

         SAMPLES <- c(rep(0, tab0[1]), S)
       }
       else SAMPLES <- rep(0,tab0[1])

       sampleMean[i] <- mean(SAMPLES) 
       sampleMedian[i] <- median(SAMPLES) 
    }
  }
}

## maeCli <- mean(abs(obs - median(obs)))
## maeEns <- mean(abs(obs - apply(ensembleData, 1, median)))

 maeCli <- mean(abs(obs - mean(obs)))
 maeEnsMean <- mean(abs(obs - apply(ensembleData, 1, mean, na.rm = TRUE)))
 maeEnsMedian <- mean(abs(obs - apply(ensembleData, 1, median, na.rm = TRUE)))

 if (is.null(nSamples)) {
   maeBMAmedian <- mean(abs(obs - Q))
   maeBMAmean <- mean(abs(obs - predictiveMean))
 }
 else {
   maeBMAmedian <- mean(abs(obs - sampleMedian))
   maeBMAmean <-   mean(abs(obs - sampleMean))
 }

## c(climatology = maeCli, ensemble = maeEns, BMA = maeBMA)

##A <- matrix( c(maeEnsMean, maeEnsMedian, maeBMAmean, maeBMAmedian), 2, 2,
##        dimnames = list(c("mean", "median"), c("ensemble", "BMA")))

##c(ensemble = A[2,1,1], BMA = A[2,2,1])
c(ensemble = maeEnsMedian, BMA = maeBMAmedian)
}